AUTONOMOUS

UNDERWATER VEHICLES
Edited by Nuno A. Cruz
Contents

Preface IX

Part 1 Vehicle Design 1

Chapter 1 Development of a Vectored Water-Jet-Based

Spherical Underwater Vehicle 3
Shuxiang Guo and Xichuan Lin

Chapter 2 Development of a Hovering-Type Intelligent

Autonomous Underwater Vehicle, P-SURO 21
Ji-Hong Li, Sung-Kook Park,
Seung-Sub Oh, Jin-Ho Suh,
Gyeong-Hwan Yoon and Myeong-Sook Baek

Chapter 3 Hydrodynamic Characteristics

of the Main Parts of a Hybrid-Driven
Underwater Glider PETREL 39
Wu Jianguo, Zhang Minge and Sun Xiujun

Part 2 Navigation and Control 65

Chapter 4 Real-Time Optimal Guidance

and Obstacle Avoidance for UMVs 67
Oleg A. Yakimenko and Sean P. Kragelund

Chapter 5 Formation Guidance

of AUVs Using Decentralized Control Functions 99
Matko Barisic, Zoran Vukic and Nikola Miskovic

Chapter 6 Modeling and Motion Control Strategy for AUV 133

Lei Wan and Fang Wang

Chapter 7 Fully Coupled 6 Degree-of-Freedom

Control of an Over-Actuated
Autonomous Underwater Vehicle 147
Matthew Kokegei, Fangpo He and Karl Sammut
VI Contents

Part 3 Mission Planning and Analysis 171

Chapter 8 Short-Range Underwater Acoustic

Communication Networks 173
Gunilla Burrowes and Jamil Y. Khan

Chapter 9 Embedded Knowledge and Autonomous Planning:

The Path Towards Permanent Presence
of Underwater Networks 199
Pedro Patrón, Emilio Miguelañez and Yvan R. Petillot

Chapter 10 Deep-Sea Fish Behavioral Responses to Underwater Vehicles:

Differences Among Vehicles, Habitats and Species 225
Franz Uiblein

Chapter 11 Mapping and Dilution Estimation

of Wastewater Discharges Based on Geostatistics
Using an Autonomous Underwater Vehicle 239
Patrícia Ramos and Nuno Abreu
Preface

Autonomous Underwater Vehicles (AUVs) are remarkable machines that

revolutionized the process of gathering ocean data. Their major breakthroughs
resulted from successful developments of complementary technologies to overcome
the challenges associated with autonomous operation in harsh environments. This
book brings together the work of many experts in several domains related to the
design, development and deployment of AUVs.

During the last decades, AUVs have gone through notable developments. In the late
eighties and early nineties, the first prototypes required a tremendous effort and
ingenious engineering solutions to compensate for the technological limitations in
terms of computational power, batteries, and navigation sensors. To deploy these
expensive vehicles navigating autonomously in a very unforgiving environment, and
expecting them to return safely was a true act of faith in engineering, a scaled version
of the early efforts in space technology.

The initial developments continued steadily and, by the end of the last century, AUVs
have gradually moved from the controlled academic environment into challenging
operational scenarios, covering scientific, commercial and military applications. As the
technology matured, many different solutions were effectively demonstrated, in
various sizes and configurations, and a few evolved into commercial products.

Underwater robotics is a peculiar field of knowledge, bringing together specific

complementary knowledge in mechanical and electrical engineering, and also in
computer science. In the last decade, with the impressive improvements in
computational power, battery technology, and miniaturization of electronic systems,
AUVs became less cumbersome and more amenable to be used as test beds for new
techniques for data processing. As smaller, lighter, and less expensive equipment
became available, the access to operational vehicles was further facilitated and more
and more prototypes became accessible for testing new algorithms and solutions. The
geographic span of valuable scientific work with field results was extended to include
a larger number of researchers, not only from leading scientific institutions but also
from more modest laboratories in emerging countries. This has resulted in an
exponential increase in AUV development and deployment, alone or in fleets, with
arguably many thousands of hours of operations accumulated around the world, and
X Preface

corresponding amount of data. Autonomous Underwater Vehicles became a common

tool for all communities involved in ocean sampling, and are now a mandatory asset
for gathering detailed ocean data at very reasonable costs.

Most of the advances in AUV capabilities aimed at reaching new application scenarios
and decreasing the cost of ocean data collection, by reducing ship time and
automating the process of data gathering with accurate geo location. Although this
yielded significant improvements in efficiency, new approaches were also envisaged
for a more productive utilization of this new tool. With the present capabilities, some
novel paradigms are already being employed to further exploit the on board
intelligence, by making decisions on line based on real time interpretation of sensor
data. In many organizations, this ability is also being applied to allow the AUVs to
conduct simple intervention tasks.

The design of Autonomous Underwater Vehicles is governed by a complex tradeoff

between the critical requirements of the planned missions, and the main constraints on
fabrication, assembly and operational logistics. Contrary to the early tendency to
develop general purpose vehicles, the current pursuit of efficiency has pushed the
concept of specific vehicles for specific tasks, frequently taking advantage of modular
designs to accelerate the assembly time.

In the last years, there have been a great number of publications related to underwater
robotics, not only in traditional engineering publications, but also in other fields where
the robotic solutions are being used as a tool to validate scientific knowledge. There
are also numerous conferences held each year, addressing all aspects of AUV
development and usage. Both have served to report the major breakthroughs and
constitute a foremost source of reference literature. This book collects a set of self
contained chapters, covering different aspects of AUV technology and applications in
more detail than is commonly found in journal and conference papers. The progress
conveyed in these chapters is inspiring, providing glimpses into what might be the
future for vehicle technology and applications.

Nuno A. Cruz
INESC Porto - Institute for Systems and
Computer Engineering of Porto
Portugal
 Part 1

Vehicle Design
 0
1

Development of a Vectored Water-Jet-Based

Spherical Underwater Vehicle
Shuxiang Guo and Xichuan Lin
Kagawa University
Japan

1. Introduction
The applications of underwater vehicles have shown a dramatic increase in recent years,
such as, mines clearing operation, feature tracking, cable or pipeline tracking and deep
ocean exploration. According to different applications, the mechanical and electrical
configuration and shape of an underwater vehicle are different. For instance, manipulators
are necessary when doing mines clearing operation or some other tasks which need to deal
with environment. If an underwater vehicle is used for underwater environment detection or
observation, it is better to make this vehicle smaller and flexible in motion that it can go to
smaller space easily. If the vehicle needs high speed moving in the water then a streamline
body is required.
Different structures with different size of underwater vehicles are developed. Most of these
underwater vehicles are torpedo-like with streamline bodies, like (Sangekar et al., 2009). And
there are some small size AUVs like (Allen et al., 2002) and (Madhan et al., 2006). And also
there are some other AUVs adopt different body shape, such as (Antonelli & Chiaverini, 2002).
Meanwhile, the propulsion system is one of the critical facts for the performance of
underwater vehicles, because it is the basis of control layers of the whole system. Propulsion
devices have variable forms, for instance, paddle wheel, poles, magneto hydrodynamic drive,
sails and oars.
Paddle wheel thrusters are the most common and traditional propulsion methods for
underwater vehicles. Usually, there are at least two thrusters installed on one underwater
vehicle, one for horizontal motion and the other for vertical motion. The disadvantages of
paddle wheel thrusters are obvious, for example, it is easy to disturb the water around the
underwater vehicles. Meanwhile, the more the paddle wheel thrusters are used, the weight,
noise and energy consumption increases.
The steering strategies of traditional underwater vehicles are changing the angular of rudders
or using differential propulsive forces of two or more than two thrusters. Of course,
there are vectored propellers being used on underwater vehicles. Reference (Cavallo et al.,
2004) and (Le Page & Holappa, 2002a) present underwater vehicles with vectored thrusters.
Reference (Duchemin et al., 2007) proposes multi-channel hall-effect thrusters which involves
vector propel and vector composition. Reference (Le Page & Holappa, 2002b) proposes an
autonomous underwater vehicle equipped with a vectored thruster. At the same time, the
design of vectoring thrusters used on aircrafts is also an example of vectored propulsion
system (Kowal, 2002), (Beal, 2004) and (Lazic & Ristanovic, 2007).
4
2 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

The purpose of this research is to develop such a kind of underwater vehicle which can adjust
its attitude freely by changing the direction of propulsive forces. Meanwhile, we would like
to make the vehicle flexible when moving in the water. Inspired by jet aircraft, we adopt
vectored water-jet propellers as the propulsion system. According to the design purpose, a
symmetrical structure would be better for our underwater vehicle (Guo et al., 2009).
This spherical underwater vehicle has many implementation fields. Because of its flexibility,
our vehicle can be used for underwater creatures observation. For example, we can install
underwater cameras on the vehicle. It can track and take photos of fishes. Another example is
that, due to its small size, we can use it to detect the inside situation of underwater oil pipes.

2. Mechanical and electrical design

2.1 Mechanical system design
Before the practical manufacture, we try to give a conceptual design of the whole structure for
this spherical underwater vehicle. At this stage, we need to consider about the dimension,
weight distribution, material, components installation, and so on. And we also need
to consider about the configuration of the propulsion system, for example, how many
water-jet propellers should we use for the purpose of optimizing power consumption without
decreasing propulsion ability. Therefore, by all of that mentioned above, we give the
conceptual designed structure of our spherical underwater vehicle as shown in Fig.1.
It adopts a spherical shape, all the components are installed inside the body. Its radius is
20cm which is smaller than that in (Antonelli et al., 2002). Its overall weight is about 6.5kg. Its
working depth is designed to 0 10m, with a max speed of about 1.5m/s.
Inside the vehicle, there will be three water-jet propellers used as propulsion system, which
is enough for surge, yaw and heave. One waterproof box is used for all the electronic
components such as sensors, batteries and the control boards. And all of these are mounted
on a triangle support which is fixed on the spherical hull. The whole structure is symmetrical
in z-axis. Therefore, it can rotate along z-axis, and by doing this, the vehicle can change its
orientation easily.

2.1.1 The spherical hull

As shown in Fig.2, the spherical hull of this underwater vehicle is made of acrylic which is
light and easy to be cut. It is about 3mm thick and the diameter is 40cm. Actually, we can
see that this spherical hull is composed of two transparent hemisphere shells. There are three
holes which can provide enough space for water-jet propellers to rotate for different motions.
We will discuss the details about the principles of the water-jet propulsion system in the next
section.

2.1.2 The waterproof box

Waterproof is essential for underwater vehicles. Fig.3 shows the design of the waterproof box.
The whole size of this box is about 22cm(hight) × 14cm(inner diameter). An O-ring is used for
seal, which has the ability to provide waterproof in our case. Inside the waterproof box, there
will be two control boards, one or two lithium batteries, depending on tasks. Meanwhile, at
the top part inside the box, there will be an digital rate gyro sensor for orientation feedback.
The body of waterproof box is also transparent, therefore, we can easily observe the inside
working status .
Development of Water-Jet-Based
Development of a Vectored a VectoredSpherical
Water-Jet-Based
Underwater Vehicle Spherical Underwater Vehicle 53

(a) Top View (b) Front View

(c) Side View (d) Computer Rendering

Fig. 1. Mechanical System Schematics of the Spherical Underwater Vehicle

(a) Design (b) Prototype

Fig. 2. Spherical Hull

(a) Design (b) Prototype

Fig. 3. Design of Waterproof Box

2.1.3 Mechanism of the water-jet propulsion system

Fig.4 is the structure of one single water-jet propeller. It is composed of one water-jet thruster
and two servo motors (above and side). The water-jet thruster is sealed inside a plastic box for
waterproof. And we use waterproof glue on servo motors for waterproof. The thruster can be
6
4 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

rotated by these two servo motors, therefore, the direction of jetted water can be changed in
X-Y plane and X-Z plane, respectively.

(a) Design (b) Prototype

Fig. 4. Structure of a Water-jet Propeller

Three of the water-jet propellers are mounted on the metal support frame, as shown in Fig.5.
Three of them are circumferentially 2π/3 apart from each other.

(a) Design (b) Prototype

Fig. 5. Water-jet Propellers mounted on Support Frame

2.2 Electrical system design

We adopt a minimal hardware configuration for the experimental prototype vehicle. For a
single spherical underwater vehicle, there are three major electrical groups, sensor group,
control group and actuator group. Fig.6 gives the electrical schematics. At present, we only
use one pressure sensor for depth control and one gyro sensor for surge control. One ARM7
based control board is used as central control, data acquisition, algorithm implement and
making strategic decisions. One AVR based board is used as the coprocessor unit for motor
control. It receives the commands from ARM and translates the commands into driving
signals for the water-jet propellers.
Fig.7 gives the main hardware for this vehicle. Fig.7(a) is the ARM7 based board with
S3C44B0X on it, which can fulfill our requirement at present. Fig.7(b) is the AVR based board
with ATmega2560 on it. RS232 bus is used for the communication between ARM7 and AVR.
In Fig.7(c) is the set of pressure sensor with the sensor body(right) and its coder (left). It use
RS422 bus for data transmission. Digital gyro sensor CRS10 is shown in Fig.7(d), we use the
build in AD converter of S3C44B0X for data acquisition.
Development of Water-Jet-Based
Development of a Vectored a VectoredSpherical
Water-Jet-Based
Underwater Vehicle Spherical Underwater Vehicle 75

Fig. 6. Electrical Schematics for Prototype System

(a) S3C44B0X Board (b) ATmega2560 Board (c) Pressure Sensor (d) Gyro Sensor

Fig. 7. Electrical Components for the Experimental Prototype Underwater Vehicle

2.3 Power supply

We adopt two power supply for the spherical underwater vehicle. The highest power
consumption components in our vehicle are propellers. For each of them, the thruster has
a working voltage of 7.2V and 3.5A current drain, servo motors can work under 5V with
relatively small current. Therefore, we use two 2-cells LiPo batteries as the power supply
for the propellers. The capacity of each battery is 5000mAh with parameter of 50c − 7.4V.
Besides, we use 4 AA rechargeable batteries for the control boards. We carried out the power
consumption test for one LiPo battery, and Fig.8 gives the battery discharge graph of the
power system.

Fig. 8. Power Consumption of the Whole System. Blue line – one propeller working; green
line – two propellers working; red line – three propellers working

3. Principles and modeling of the propulsion system

In this section, we will discuss about the working principles, modeling method and the
identification experiment for the water-jet propeller. Many literatures have presented the
computing formula for the torque and thrust exerted by a thruster. Most of them are base on
8
6 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

the lift theory, and mainly focus on blades type propellers (Newman, 1977), (Fossen, 1995) and
(Blanke et al., 2000). Our propellers are different with blades type propellers, therefore, we try
to find another method for the modeling of water-jet propellers. In (Kim & Chung, 2006), the
author presented a dynamic modeling method in which the flow velocity and incoming angle
are taken into account. We will use this modeling method for our water-jet propellers.

3.1 Working principles

Before modeling of propulsion system, we want to give some basic working principles about
the water-jet propellers. Fig.9(a) shows the top view of distribution of three propellers. They
can work together to realize different motion, such as surge and yaw.

(a) Propeller Distribution (b) propeller-fixed Coordinates

Fig. 9. Distribution and Coordination of Multiple Propellers

If we let θ be the interval angle of each water-jet propeller, as shown in Fig.9(b), then, for
the purpose of kinematics transform, three propeller-fixed coordinates are introduced for
propellers, which are fixed in the rotation center of the propellers. So we can see, these three
propeller-fixed coordinates are actually transform results of vehicle-fixed coordinate reference
frame. Meanwhile, it should be noted that, this transform only happens in X-Y plane. Let the
matrix form of the coordinates transform be given as:
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
X1 X −R
⎝ Y1 ⎠ = ⎝ Y ⎠ + ⎝ 0 ⎠ (1)
Z1 Z 0

⎛ ⎞
⎛ ⎞ ⎛ ⎞⎛ ⎞ 1 π
X2 cθ sθ 0 X ⎜ Rcθ − Rsθc ⎟
⎜ 2 6 ⎟
⎝ Y2 ⎠ = ⎝ − sθ cθ 0 ⎠⎝ Y ⎠+⎜ 1 π ⎟ (2)
⎝ − Rcθ − Rcθc ⎠
Z2 0 0 1 Z 2 6
0

⎛ ⎞
⎛ ⎞ ⎛ ⎞⎛ ⎞ 1 π
X3 c2θ s2θ 0 X ⎜ Rc2θ + Rs2θc ⎟
⎜ 2 6 ⎟
⎝ Y3 ⎠ = ⎝ − s2θ c2θ 0 ⎠⎝ Y ⎠+⎜ 1 π ⎟ (3)
⎝ − Rc2θ + Rc2θc ⎠
Z3 0 0 1 Z 2 6
0

where R is the radius of the vehicle, s(·) ≡ sin(·) and c(·) ≡ cos(·).
Development of Water-Jet-Based
Development of a Vectored a VectoredSpherical
Water-Jet-Based
Underwater Vehicle Spherical Underwater Vehicle 97

So, a general transform matrix can be obtained:

p
Pb = Φbp · p P p + C (4)

where p Pb is the position vector of propeller-fixed coordinate expressed in vehicle-fixed

coordinate, Φbp = (Φbp1 , Φbp2 , Φ bp3 ) T is the transform matrix from propeller-fixed coordinate
to vehicle-fixed coordinate, p P p is the position vector in propeller-fixed coordinate and the C
is a constant vector.
Now, let us take a look at three motions, surge, heave and yaw. The definition of these
three motions can be found in (Fossen, 1995). Before that, we define two angles which will
be used for orientation of propellers. Fig.10 gives the definition of α and β. Fig.11 gives a
demonstration of surge, heave and yaw.

(a) Rotation in X-Y Plane (b) Rotation in X-Z Plane

Fig. 10. Orientation of Propellers

(a) Surge (b) Heave (c) Yaw

Fig. 11. Propulsion Forces for Surge, Heave and Yaw

The first case is surge. In this case, two of the water-jet propellers will work together, and the
other one could be used for brake. So, from Fig.11(a), two water-jet propellers in the left will
be used for propulsion, and if we want to stop the vehicle from moving, the third propeller
can act as a braking propeller. From Equation 4, the resultant force for surge can be expressed
in vehicle-fixed coordinate as:
⎧
⎪ T 3
⎪
⎨ p Fxb = Φbp1 ∑ ( p F ip + e1 Ci ) = 0
i =1 (5)
⎪ pF = 0
⎪
⎩ p yb
Fzb = 0
10
8 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

where, e1 = (1, 0, 0) T .
Then, for the heave case, all the three water-jet propellers will work and the side servo motor
will rotate to an angle that β > π/2. Therefore, in this case, the resultant force for heave can
be expressed in vehicle-fixed coordinate as:
⎧ p
⎪
⎪ Fxb = 0
⎨ pF = 0
yb
(6)
⎪
⎪ T 3
⎩ Fzb = Φbp3 ∑ ( p F ip + e3 Ci ) = 0
p
i =1

where, e3 = (0, 0, 1) T .
The third case is yaw which is rotating on z-axis. By denoting in propeller-fixed coordinates,
α should have the same orientation, clockwise or counterclockwise, that means, αi > 0 or
αi < 0. So in yaw, rotation moment will take effect. We can write the equation for yaw in
vehicle-fixed coordinate as:
⎧ 3
⎪
⎪ p F = Φb T
∑ ( p F ip + e1 Ci ) = 0
⎪
⎪ xb p1
⎪
⎪ i =1
⎨ 3
p F = Φb T
yb ∑ ( p F ip + e2 Ci ) = 0 (7)
⎪
⎪
p2
i =1
⎪
⎪ pF = 0
⎪
⎪
⎩ p zb
M xb + p Myb + p Mzb = 0

where, e2 = (0, 1, 0) T , e3 = (0, 0, 1) T .

3.2 Modeling of single water-jet propeller

In the author’s previous research (Guo et al., 2009), the modeling for orientation of water-jet
propeller is presented. Therefore, in this part, we will only discuss about the hydrodynamics
modeling of the water-jet thruster. The method we refer to is presented in (Kim & Chung,
2006). For the purpose of dynamic modeling of water-jet propeller, we give the flow model of
the water-jet thruster, which is shown in Fig.12. The shaft is perpendicular to the nozzle, and
there are two blades.

Fig. 12. Flow Model of the Water-jet Thruster (top view)

where,
Ω is angular velocity of the thruster
Vi is velocity of incoming flow
Vc is central flow velocity in the nozzle
Development of Water-Jet-Based
Development of a Vectored a VectoredSpherical
Water-Jet-Based
Underwater Vehicle Spherical Underwater Vehicle 119

D is diameter of the nozzle

Vo is velocity of outlet flow
γ is incoming angle of ambient flow
Because the diameter of the nozzle is small, the velocity difference in the nozzle can be
ignored, so we consider the axis flow velocity Va as a linear combine of incoming flow velocity
and the central flow velocity,

Va = k1 Vi + k2 Vc
1
Vc = DΩ (8)
2
Vi = Vf cosγ

By assuming that the flow is incompressible, therefore, from equation of continuity, we know
that the volume of incoming flow must equal to the outlet flow, then we get:

ρ a Va A a = ρo Vo Ao (9)

where, ρ a = ρo is density of flow, A a = Ao is cross-section of the nozzle.

Therefore, we can also get:

Va = Vo (10)

Meanwhile, we know that, the propulsive force of the water-jet thruster is:

Ft = ρAV 2a (11)

By substituting Equation 8 in Equation 11, we can get:

π
Ft = ρD2 (k21 Vi 2 + 2k1 k2 Vi DΩ + k22 D2 Ω2 ) (12)
4
By rewriting Equation 12, we get:

Ft π V V
= (k21 ( i )2 + 2k1 k2 i + k22 ) (13)
ρD4 Ω2 4 DΩ DΩ

Then, we can let the non-dimensional parameter be:

π 2 Vi 2 V
K T ( J0 ) = ( k1 ( ) + 2k1 k2 i + k22 ) (14)
4 DΩ DΩ
where
Vi Vf cosγ
J0 = = (15)
DΩ DΩ
J0 is the advance ratio.
Now, the modeling becomes measuring of three parameters, flow velocity, incoming angle
and angular velocity of thruster. For this purpose, we designed an experiment to measure
these parameters and find out their relationship.
12
10 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

Flow velocities Depth Incoming angles Control voltages

0.1m/s
80cm 0−π 3V − 7V (DC)
0.2m/s

Table 1. Experiment Condition

3.3 Experiments for the dynamics modeling

In this part, we try to identify the dynamics model of the water-jet propeller by experiment.
What we are interested in is the relation of flow incoming angles, flow velocities and
propulsive forces. Experiment condition is listed in Table 1.

3.3.1 Experiment design

Fig.13 gives the experiment principle. We use one NEC 2301 stain gage as the force sensor and
use NEC AC AMPLIFIER AS 1302 to amplify the output signal from strain gage, which are
shown in Fig.14.

Fig. 13. Experiment Design for Identification

Fig. 14. Strain Gage and Amplifier

Firstly, we give a brief illustration for the strain gage measurement. Let Fd be deformation
force, and ε be the deformation of aluminium lever used in the experiment. Therefore, from
the theory of mechanics of materials, we can get the relation of Fd and ε as:
ZE
Fd = ε (16)
X
where, Z is second moment of area, E is the Young’s modulus, X is the distance from acting
point of force to stain gage.

3.3.2 Experiments and analysis

In the experiment, there are four variables we need to consider, the equivalent cross-section of
propeller, flow velocity, incoming angle and control voltage. What we are interested in is the
variation of propulsive force in different incoming angles and different control voltage.
Development of Water-Jet-Based
Development of a Vectored a VectoredSpherical
Water-Jet-Based
Underwater Vehicle Spherical Underwater Vehicle 13
11

3.3.2.1 Equivalent cross-section variation of propellers

As a vectored water-jet-based propulsion system, it should be noted that both the propulsive
force and its direction can be changed. Therefore, when the propeller changes its direction,
actually, the incoming angle of flow is also changing, and the equivalent cross-section of
the propeller is changing. From Equation 11 we know the propulsive force will change if
cross-section A changes. Fig.15 gives a demonstration of this case. When the propeller
rotate from position I to II, the equivalent cross-section will change from cross-section I to
cross-section II. So we try to find an equation to describe this variation.

Fig. 15. Variation of Equivalent Cross-section of Propellers

Considering that the measured force from stain gage is actually a resultant force of propulsive
force and fluid force. And we also know that the fluid force acted on the propeller depends
on the equivalent cross-section.
So the first experiment is measurement of the equivalent cross-section variation. The propeller
is submerged in the flow which has a speed of 0.2m/s, propeller is powered off. And we only
change its orientation in X − Y plane. Because of experiment limits, we can not change the
flow direction, so in the experiment, the incoming angle equals to the orientation angle of the
propeller. Fig.15 gives a demonstration of the equivalent cross-section. We give some special
angles, 0, π/6, π/3, 2π/3, 5π/6, π, for this experiment. Fig.16 gives the experiment data of
equivalent cross-section. You may notice that, we did not adopt the orientation angle of π/2.
Because, when the propeller rotate to π/2, which means that the measure surface of the strain
gage is parallel to the flow direction, the strain gage can not measure the flow force.

Fig. 16. Variation of Equivalent Cross-section

14
12 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

From Fig.16 we can see, the data curve is similar with a sinusoid, so we use a sine function to
fit this experiment data:
Ae (φ) = λ1 sin (λ2 φ + λ3 ) (17)
where φ is incoming angle, λ1 , λ2 , λ3 are coefficients.
3.3.2.2 Incoming angle and deformation force
In this case, the flow velocity is seen as constant. Two groups of experiment are carried out at
flow velocity of 0.1m/s and 0.2m/s. The control voltage to thruster is from 3V to 7V every 1V.
From the data shown in Fig.17, we can see that the deformation force does not simply increase

(a) Vf = 0.1m/s

(b) Vf = 0.2m/s

Fig. 17. Deformation Forces with Different Incoming Angles

with the increasing of incoming angle, the maximum deformation of the lever happens at
about 60 degree of the incoming angles. They are not a linear relation. Then, how about the
real propulsive force?
3.3.2.3 Incoming angle and propulsive force
As we mentioned, what we measured by stain gage is actually a resultant force of propulsive
force and fluid force. Therefore, the real propulsive force Ft should be calculated using
deformation force Fd and the equivalent cross-section Ae .
Fd = Ft − F f cosγ (18)
If we consider about the fluid force of the flow, we can refer to Equation 11 and 17, then write
the fluid force as:
F f = ρV 2c Ae (φ) (19)
Development of Water-Jet-Based
Development of a Vectored a VectoredSpherical
Water-Jet-Based
Underwater Vehicle Spherical Underwater Vehicle 15
13

We substitute Equation 19 into 18 we get:

Ft = Fd + ρV 2c Ae (φ)cosγ (20)

So now, we can calculate the real propulsive force by using equivalent cross-section,
deformation force and incoming angles. The results is shown in Fig.18. Fig.18(a) and Fig.18(b)
are results at the flow velocity of Vf = 0.1m/s and Vf = 0.2m/s, respectively.

(a) Vf = 0.1m/s

(b) Vf = 0.2m/s

Fig. 18. Propulsive forces with Different Incoming Angles

3.3.2.4 Control voltage and propulsive force

From the results of 3.3.2.3, we can obtain the relation of control voltage and propulsive force.
First, we give the experiment data, in Fig.19. From the diagram, we can see that the relation
of control voltage and propulsive force can be described using linear equation.

4. Underwater experiments for basic motions

Because of the symmetrical shape of the hull, it is obvious that motion characteristics of surge,
sway and heave should be similar. However, from another point of view, surge and sway are
motions in X − Y plane while heave is a motion that its motion surface perpendicular to X − Y
plane. Therefore, we carry out experiments for horizontal motion surface and vertical motion
surface respectively. Besides, for the experimental prototype vehicle, we only consider one
rotational DOF in Z axis, so the third experiment is yaw motion.
16
14 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

(a) Vf = 0.1m/s

(b) Vf = 0.2m/s

Fig. 19. Propulsive forces with Different Control Voltages

4.1 Experiment of horizontal motion

This experiment combines surge and sway together to verify the motion characteristics of the
vehicle in horizontal plane. We carried out three experiments:
Case 1:
step 1. Surge (Move forward in X axis);
step 2. Right steering (Turn right about 90o );
step 3. Sway (Move forward along Y axis.)
Case 2:
step 1. Surge (Move forward in X axis);
step 2. Left steering (Turn left about 90o );
step 3. Sway (Move forward along Y axis.)
Case 3:
step 1. Surge (Propeller I and II work together, propeller III powered off);
step 2. Brake (Propeller I and II powered off, Propeller III works to produce brake force).
In case 1, timing of step 1 is about 10s, and step 2 takes about 12s. And timing of case 2 is
relatively the same with case 1, because of the same hydrodynamics characteristics of turning
right and left. In case 3, it takes 15s reaching a stable speed, and the brake effect happens
in about 3s which is effective for low speed underwater vehicles. From Fig.20, we can see,
Development of Water-Jet-Based
Development of a Vectored a VectoredSpherical
Water-Jet-Based
Underwater Vehicle Spherical Underwater Vehicle 17
15

(a) The trajectory of case 1 (b) Steering angle of case 1

(c) The trajectory of case 2 (d) Steering angle of case 2

(e) Surge and Brake of case 3

Fig. 20. Experimental Results of Horizontal Motion

the experimental results fit well with simulation results in surge stage, but when the vehicle
rotating, errors become large. The reason of this is because the simulation experiment only
considered linear damping force and quadratic damping force, but in reality, there are other
hydrodynamic forces act on the vehicle.

4.2 Experiment of vertical motion

Even though we design the working depth of the vehicle to 10m, because the depth of
experimental pool is only 1.2m, we can only make the experiments in shallow water. So we
set the vertical motion time in a relatively small range. We also carried out two experiments:
Case 1:
step 1. Set the top point of the spherical hull as the start point;
step 2. Move downward in Z axis for about 7s ;
step 3. Float up to the surface.
Case 2:
step 1. Set the top point of the spherical hull as the start point;
18
16 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

step 2. Move downward in Z axis for about 7s ;

step 3. Stop the vehicle.

(a) Submerge and Float up (b) Submerge Only

Fig. 21. Experimental Results of Vertical Motion

From Fig.21(a) and Fig.21(b), we can see, the experimental results does not fit well with
simulation results very well, errors exceed 100%. When we analyze the reasons, we find that,
the simulation experiment does not consider the variation of water pressure. The control
voltage to the thrusters is 7V as a constant. That means, the propulsive force will not change.
But with the increasing of depth, water pressure increases. As a result, the effective propulsive
force are weaken by water pressure.

4.3 Experiment of yaw

We let the vehicle rotate about 90o then stop. From Fig.22(a) and Fig.22(b), the maximum error
between simulation results and experimental results happens at about 2.8s where is nearly
the maximum angular velocity. The reason of this result is that, we simplified the model of
our vehicle, especially the hydrodynamic damping forces. Only linear damping force and
quadratic damping force are taken into account in our case. But in the real experiment, there
are many other velocity related hydrodynamic damping forces, therefore, when the angular
velocity increasing, the damping effect of ignored forces become obvious.

(a) Angular Velocity of Yaw (b) Angle of Yaw

Fig. 22. Experimental Results of Yaw

5. Conclusions
In this paper, we proposed a spherical underwater vehicle which uses three water-jet
propellers as its propulsion system. We introduced the design details of mechanical and
electrical system.
Development of Water-Jet-Based
Development of a Vectored a VectoredSpherical
Water-Jet-Based
Underwater Vehicle Spherical Underwater Vehicle 19
17

Based on the design of the vehicle, we introduced the principles of the water-jet propulsion
system including the force distribution of three water-jet propellers, the working principles
of different motions. And then we discussed about the modeling of one single propeller by
identification experiments. For the modeling, the flow velocity and equivalent cross-section
of the propeller are taken into account for dynamics model.
One experimental prototype of this spherical underwater vehicle is developed for the purpose
of evaluation. Underwater experiments are carried out to evaluate the motion characteristics
of this spherical underwater vehicle. Experimental results are given for each experiment, and
the analysis are also given.
From the underwater experiments of the prototype vehicle, the availability of the design is
proved, and the water-jet propulsion system can work well for different motions. But there
are also some problems needed to be resolved. Firstly, the propulsive force of the water-jet
propellers needed to be increased; secondly, the variation of water pressure on the propulsive
force should be considered when building the dynamics model of propellers; thirdly, the
gravity distribution should be re-regulated to improve stability; finally, from experiments, it
is necessary to improve the accuracy of the dynamics model of the vehicle for precise control.

6. References
Allen, B., Stokey, R., Austin, T., Forrester, N., Goldsborough, R., Purcell, M. & von Alt,
C. (2002). REMUS: a small, low cost AUV; system description, field trials and
performance results, OCEANS’97. MTS/IEEE Conference Proceedings, Vol. 2, IEEE,
pp. 994–1000.
Antonelli, G. & Chiaverini, S. (2002). Adaptive tracking control of underwater
vehicle-manipulator systems, Proceedings of the 1998 IEEE International Conference on
Control Applications, 1998, Vol. 2, IEEE, pp. 1089–1093.
Antonelli, G., Chiaverini, S., Sarkar, N. & West, M. (2002). Adaptive control of an autonomous
underwater vehicle: experimental results on ODIN, IEEE Transactions on Control
Systems Technology 9(5): 756–765.
Beal, B. (2004). Clustering of Hall effect thrusters for high-power electric propulsion applications,
PhD thesis, The University of Michigan.
Blanke, M., Lindegaard, K. & Fossen, T. (2000). Dynamic model for thrust generation of marine
propellers, Proceedings of the IFAC Conference on Maneuvering and Control of Marine
Craft (MCMC 2000), Citeseer.
Cavallo, E., Michelini, R. & Filaretov, V. (2004). Conceptual design of an AUV Equipped with
a three degrees of freedom vectored thruster, Journal of Intelligent & Robotic Systems
39(4): 365–391.
Duchemin, O., Lorand, A., Notarianni, M., Valentian, D. & Chesta, E. (2007). Multi-Channel
Hall-Effect Thrusters: Mission Applications and Architecture Trade-Offs, 30th
International Electric Propulsion Conference, Florence, Italy.
Fossen, T. I. (1995). Guidance and Control of Ocean Vehicles, John Wiley & Sons Ltd., USA.
Guo, S., Lin, X. & Hata, S. (2009). A conceptual design of vectored water-jet propulsion
system, International Conferenceon Mechatronics and Automation, 2009. ICMA 2009,
IEEE, pp. 1190–1195.
Kim, J. & Chung, W. (2006). Accurate and practical thruster modeling for underwater vehicles,
Ocean Engineering 33(5-6): 566–586.
Kowal, H. (2002). Advances in thrust vectoring and the application of flow-control technology,
Canadian aeronautics and space journal 48(2): 145–151.
20
18 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

Lazic, D. & Ristanovic, M. (2007). Electrohydraulic thrust vector control of twin rocket
engines with position feedback via angular transducers, Control Engineering Practice
15(5): 583–594.
Le Page, Y. & Holappa, K. (2002a). Hydrodynamics of an autonomous underwater
vehicle equipped with a vectored thruster, OCEANS 2000 MTS/IEEE Conference and
Exhibition, Vol. 3, IEEE, pp. 2135–2140.
Le Page, Y. & Holappa, K. (2002b). Simulation and control of an autonomous underwater
vehicle equipped with a vectored thruster, OCEANS 2000 MTS/IEEE Conference and
Exhibition, Vol. 3, IEEE, pp. 2129–2134.
Madhan, R., Desa, E., Prabhudesai, S., Sebastião, L., Pascoal, A., Desa, E., Mascarenhas,
A., Maurya, P., Navelkar, G., Afzulpurkar, S. et al. (2006). Mechanical design and
development aspects of a small AUV–Maya, 7th IFAC Conference MCMC2006.
Newman, J. (1977). Marine hydrodynamics, The MIT press.
Sangekar, M., Chitre, M. & Koay, T. (2009). Hardware architecture for a modular autonomous
underwater vehicle STARFISH, OCEANS 2008, IEEE, pp. 1–8.
 2

Development of a Hovering-Type Intelligent

Autonomous Underwater Vehicle, P-SURO
Ji-Hong Li1*, Sung-Kook Park1, Seung-Sub Oh1, Jin-Ho Suh1,
Gyeong-Hwan Yoon2 and Myeong-Sook Baek2
1Pohang Institute of Intelligent Robotics
2Daeyang Electric Inc.

Republic of Korea

1. Introduction
P-SURO(PIRO-Smart Underwater RObot) is a hovering-type test-bed autonomous
underwater vehicle (AUV) for developing various underwater core technologies (Li et al.,
2010). Compared to the relatively mature torpedo-type AUV technologies (Prestero, 2001;
Marthiniussen et al., 2004), few commercial hovering-type AUVs have been presented so far.
This is partly because some of underwater missions of hovering-type AUV can be carried
out through ROV (Remotely Operated Vehicle) system. But the most important reason is of
less mature core technologies for hovering-type AUVs. To carry out its underwater task,
hovering-type AUV may need capable of accurate underwater localization, obstacle
avoidance, flexible manoeuvrability, and so on. On the other hand, because of limitation of
present underwater communication bandwidth, high autonomy of an AUV has become one
of basic function for hovering AUVs (Li et al., 2010).
As a test-bed AUV, P-SURO has been constructed to develop various underwater core
technologies, such as underwater vision, SLAM, and vehicle guidance & control. There are
four thrusters mounted to steer the vehicle's underwater motion: two vertical thrusters for
up/down in the vertical plane, and 3DOF horizontal motion is controlled by two horizontal
ones, see Fig. 1. Three communication channels are designed between the vehicle and the
surface control unit. Ethernet cable is used in the early steps of development and
program/file upload and download. On the surface, RF channel is used to exchange
information and user commands, while acoustic channel (ATM: Acoustic Telemetry
Modem) is used in the under water. A colour camera is mounted at the vehicle's nose. And
three range sonar, each of forward, backward and downward, are designed to assist
vehicle's navigation as well as obstacle avoidance and SLAM. An AHRS combined with 1-
axis Gyro, 1-axis accelerometer, depth sensor consist of vehicle's navigation system.
In this chapter, we report the details of to date development of the vehicle, including SLAM,
obstacle detection/path planning, and some of vehicle control algorithms. The remainder of
this chapter is organized as follows. In Section II, we introduce the vehicle's general
specifications and some of its features. Underwater vision for P-SURO AUV is discussed in
Section III, and the SLAM algorithm in the basin environment is presented in Section IV. In
Section V, we discuss some of control issues for P-SURO AUV. Finally in Section VI, we
make a brief summary of the report and some future research issues are also discussed.
22 Autonomous Underwater Vehicles

Fig. 1. P-SURO AUV and its open frame.

2. P-SURO AUV overview

As aforementioned, P-SURO AUV is a test-bed for developing underwater technologies.
And most of its experimental tests will be carried out in an engineering basin in the PIRO
with dimension of 12(L)×8(W)×6(D)m. Under these considerations, the vehicle is designed to
be compact size with easiness of various algorithm tests (Li et al., 2010). The general
specification of the vehicle is as Table. 1.

Item Specifications
Depth rating 100m
Weight 53kg
Dimension 1.05(L)×0.5(W)×0.3(H)m
Max. speed FW: 2.5knot; BW, UP/DW: 1.5knot
Battery system 400W·hr, Lithium Ion, Endurance: 2.5hrs
Payload ≤4kg

Table 1. General specification of P-SURO AUV.

2.1 Mechanical system

For the convenience of maintenance and also under the security consideration, we separate
the battery system from other electronics systems, see Fig. 2. Main frame is made of AL-
6061, fixing parts for camera and range sonar are made of POM. To increase the
hydrodynamic mobility in the underwater horizontal plane, the open frame of vehicle is
wrapped in a two-piece of FRP (Fibre-Reinforced Plastic) shell (Li et al., 2010).
Throughout its underwater missions, P-SURO is always keeping zero pitch angle using two
vertical thrusters. With this kind of stability in its pitch dynamics, the vehicle's horizontal
3DOF motion is steered by two horizontal thrusters. From control point of view, this is a
typical underactuated system. And how to design path tracking or following scheme for this
kind of underactuated system has become one of most intense research area in the nonlinear
control community (Jiang, 2002; Do et al., 2004; Li et al., 2008b).
Development of a Hovering-Type Intelligent Autonomous Underwater Vehicle, P-SURO 23

Fig. 2. Mechanical arrangement of P-SURO AUV.

Sensor Model (Maker) Specifications

Super SeaSpy - >480 TV lines
(Tritech) - 1/3" Interline Transfer CCD
- Composite output
Micron Echo Sounder - Operating frequency: 500KHz
(Tritech) - Beamwidth: 6o conical
- Operating range: 50m
AHRS - Update rate: 1-100Hz
(Innalabs) - Heading accuracy: 1.0 deg
- Attitude accuracy: <0.4 deg
CVG25 - Measurement range: ±200deg/sec
(Innalabs) - Bandwidth: 50Hz
- Bias stability: 1.5deg/hr
AL-15M2.5 - Input range: ±2.5g
(Innalabs) - Bias: <4mg
- Bandwidth: 300Hz
PR-36XW - Depth rating: 100m
(Keller) - Accuracy: 0.1% FS
- Output rate: 100Hz
UWM2000H - Payload data rate: 300-1200bits/sec
(LinkQuest) - Working range: 1200m
- Beam width: 210deg (omni-directional)
LinkWiserTM-HP400 - Half-duplex
(Cellution) - Operating frequency: 400-470MHz
- Air data rate: 4.8kbps
BTD-156 - Continual thrust: 2.2kgf
(SeaBotix) - Input voltage: 28V
- Interface: RS485, 115200bps
Table 2. Sensor & thrust system of P-SURO AUV.
24 Autonomous Underwater Vehicles

2.2 Sensor, thrust, and power system

For underwater vision, there is one colour camera mounted at the vehicle nose. And three
range sonar (forward, backward and downward) are mounted on the vehicle. There sonar
are designed for obstacle detection and also for assisting vehicle's underwater localization.
For P-SURO AUV, we design a relatively simple but low grade of inertial navigation system
which consists of AHRS, 1-axis Gyro, 1-axis accelerometer, one depth sensor.
SeaBotix BTD-156 thrusters are selected to steer the vehicle's underwater motion. This is a
small size underwater thruster with 90W of average power consumption. For power system,
the calculated total power consumption of vehicle system is about 450W. And
correspondingly, we design the 1.2kW Lithium Ion battery system, which can support more
than two hours of the vehicle's underwater continuous operation. The overall sensor &
thrust system for P-SURO AUV is listed in Table. 2.

2.3 Embedded system

Three of PC104 type PCM3353 SBCs (Single Board Computers) are chosen as core modules,
each of vision, navigation, and control. PCM3353 SBC provides 4 RS232 channels plus 4 USB
channels. And using these USB channels, we can easily extend the necessary serial channels
(RS232/422/485) using proper USB to serial converters. PCM3718HG analogue and digital
I/O board is used for various peripheral interface. In addition, two peripheral boards,
including DC/DC converter system, magnetic switch circuit, leakage detection circuit, are
also designed. Fig. 3 shows the inner view of electronic pressure hull.

Fig. 3. Electronics system of P-SURO AUV.

3. Software architecture
As aforementioned, we choose Windows Embedded CE 6.0 as the near real-time OS for
three of core modules; vision module, navigation module, and control module. For this, we
design three different WinCE 6.0 BSPs (Board Support Package) for each of three core
modules. Furthermore, these three core modules are connected to each other through
Ethernet channel, and constructing a star topology of network structure.
Development of a Hovering-Type Intelligent Autonomous Underwater Vehicle, P-SURO 25

Software frame for each core module consists of thread-based multi tasking structure. For
each module, there are various sensors connected through serial and analogue channels.
And these serial sensors, according to their accessing mechanism, can be classified into two
types: active sensor (frequently output measurement) and passive sensor (trigger mode). For
these passive sensors as well as analogue sensors, we read the measurements through
Timer( ) routine. And for each of active sensors, we design a corresponding thread. In most
of time, this thread is in Blocking mode until there is measurement output. And this kind of
real-time sensor interface also can be used to trigger other algorithm threads. For example,
in the navigation module, there is a thread designed for interfacing with AHRS sensor
(100kHz of output rate). After accessing each of attitudes, gyro, and accelerometer output
measurement, the thread will trigger Navigation( ) thread. Moreover, some of these threads
are cautiously set with different priority values.
As with the most of other AUVs so far, the P-SURO AUV has the similar overall software
frame, which can be divided into two parts: surface remote control system and the vehicle
software system. For surface system, the main functions of it are to monitor the vehicle and
deliver the user command. According to the user command (mission command in this case),
the vehicle will plan a series of tasks to accomplish the mission. For P-SURO AUV, its most
experimental field is in a small cuboid. In this kind of environment, it is well known that
underwater acoustic channel is vulnerable. For this reason, the vehicle is required to possess
relatively high level of autonomy, such as autonomous navigation, obstacle avoidance, path
planning and so on.
From the control architecture point of view, the software architecture of P-SURO AUV can
be classified into hybrid architecture (Simon et al., 1993; Healey et al., 1996; Quek & Wahab,
2000), which is a certain combinaiton of hierarchical architecture (Wang et al., 1993; Peuch et
al., 1994; Li et al., 2005) and behavioral architecture (Brooks, 1986; Zheng, 1992; Bennett,
2000). As aforementioned, because of the limitation of underwater acoustic communication
in the engineering basin in PIRO, it is strongly recommended for the vehicle to self-
accomplish its mission without any of user interface in the water. For this consideration, the
control architecture of P-SURO AUV is featured as a behavioral architecture based hybrid
system (see Fig. 4).

Fig. 4. Hybrid control architecture for P-SURO AUV.

26 Autonomous Underwater Vehicles

If there is a pattern appeared in a certain area in front of the vehicle, the vision module will
recognize the pattern and transmit the corresponding vehicle's pose information freqeuntly
to the control module for aiding of path planning. According to the received mission
command (user command is usually delivered to the vehicle on the surface through RF
channel), the control module arranges a series of tasks to accomplish the mission. Also, this
module carries out various thruster controls and other actuator controls. The main task of
the navigation module is to carry out the real-time navigatin algorithm using acquired
attitude, gyro, and accelerometer measurements. Other information including range sonar,
depth sensor, underwater vision are served as aiding information for this inertial navigatin
system.

4. Vision-based underwater localization

Visual localization methods usually can be classified into two types: one is based on natural
feature points around the environment for recognition of robot pose, and the other one is
using artificial landmarks which are usually known patterns pre-installed in the
environment (Oh et al., 2010). PIRO engineering basin is surrounded by flat concrete walls,
and it is difficult to extract specific feature points. For this reason, we use artificial landmark
for visual localization of P-SURO.

4.1 Artificial landmarks

For P-SURO AUV, the main purpose of underwater vision is to provide localization
information for the vehicle's path planning task. In the path decades, vision has become one
of most intense research area in the robot community, and various artificial landmarks and
corresponding recognition methods have been developed (Hartley & Zisserman, 2000). The
pattern designed for P-SURO AUV is shown in Fig. 5, which consists of two rectangles and
two sets of four dots with the same cross-ratio. The large number in the centre is used for
distinguish the pattern (Li et al., 2010).
The eight dots in the pattern contain 3D pose information. For extracting these dots, we use
the cross-ratio invariant, which is a basic projective invariant of perspective space (Hartley
& Zisserman, 2000). The cross-ratio is defined as following
| || |
( , , , )= , (1)
| || |

where

= .

Fig. 5. Designed underwater pattern for P-SURO AUV.

Development of a Hovering-Type Intelligent Autonomous Underwater Vehicle, P-SURO 27

Fig. 6. Cross-ratio in the projective space.

The eight dots in the pattern contain 3D pose information. For extracting these dots, we use
the cross-ratio invariant, which is a basic projective invariant of perspective space (Hartley
& Zisserman, 2000). The cross-ratio is defined as following
| || |
( , , , )= , (2)
| || |

where

= .

The cross-ratio value defined in (2) is invariant under any projective transformation of the
line. If = × , then ( ′ , ′ , ′ , ′ )= ( , , , ). As shown in Fig. 6,
suppose the plane π is the pattern and the plane π′ is the projected pattern image, then four
dots on the line have the same cross-ratio with the four dots on the line ′. Using this
invariant, we can find the match between pattern dots and its image projection.

4.2 Auto-calibration of underwater camera

It is difficult to directly calibrate the camera in the underwater environment. For this reason,
we apply a camera auto-calibration method using cross-ratio invariant (Zhang, 2000).
If we denote the pattern point as = [ , , 1] and the image point as ′ = [ ′, ′, 1] , then
the relationship between them is given by

′= , (3)
where is an arbitrary scale factor. In (3), homography is defined as

= [ℎ ℎ ℎ ] = [ ], (4)
where is a scale factor, , are part of rotation matrix = [ ], is a translation
matrix and is a camera intrinsic matrix. If camera moves, each matching point from the
scene makes a homography. We can get the camera intrinsic matrix from homography
(Zhang, 2000).
28 Autonomous Underwater Vehicles

Given the camera intrinsic matrix and the homography in the image, we can get the
three-dimensional relationship between the pattern and the camera (robot) using following
equations

= ℎ
= ℎ ,
(5)
= × ,
= ℎ .

4.3 Lab-based evaluation

To evaluate the developed vision algorithm, we carry out a series of ground tests. Fig. 7
shows the test results. First, eight dots in the pattern are extracted from the image (Fig. 7-a).
And Fig. 7-b shows the selected images for camera auto-calibration, and extracted camera
pose is shown in Fig. 7-c,d.
While evaluating the performance of proposed visual localization method, we mainly
consider two points: one is the pattern recognition rate, and the other one is the accuracy of
pose estimation. For pattern recognition rate, we arbitrarily move the camera and take 306
images. 150 of them are correctly recognized, 85 are failed to recognize, 61 are blurred
because of camera movement, and 10 do not include the full pattern. Except 61 blurred
images and 10 of missed-pattern images, the recognition rate is about 64%. However,
consider the fact that about half of the non-recognized images are rotated more than 40
degrees from the pattern, the recognition rate is about 81%.
To evaluate the accuracy of pose estimation, we fix the camera and locate the pattern at 12
known positions between 400mm to 700mm distance. Calculated average pose estimation
error is 0.155mm and the standard deviation is 1.798mm.

Fig. 7. Process of pose estimation.

4.4 Camera underwater test

Given a pattern (landmark) and a camera, then the minimum and maximum pattern
recognition range can be predetermined. And this information will be used for vehicle's
Development of a Hovering-Type Intelligent Autonomous Underwater Vehicle, P-SURO 29

underwater path planning. Under this consideration, we carry out a series of test measuring
the minimum and maximum recognition range both in the air and in the water. Test results
are shown in Fig. 8 and 9, from which we can see that the maximum recognition range in the
water is approximately half of the one in the air. For the safety consideration, we force the
vehicle to keep from the basin wall at least 1.5m throughout the various basin tests.

Fig. 8. Test environments.

Fig. 9. Test results

5. SLAM using range sonar array

In the past decades, SLAM (Simultaneous Localization and Mapping) has been one of most
hot issues in the robotics community (Leonard & Burrant-Whyte, 1992; Castellanos &
Tardos, 1999; Thrun et al., 2005). SLAM problems arise when the robot does not have access
to a map of the environment; nor does it have access to its own poses (Thrun et al., 2005).
30 Autonomous Underwater Vehicles

For P-SURO AUV, as aforementioned, most of its underwater operations are carried out in
the engineering basin in PIRO. For this reason, we have designed a relatively simple SLAM
method with partially known environment.

5.1 Range sonar model

There are three Micron Echosounder@Tritech (6o conical beamwidth with 500kHz operating
frequency) mounted on the vehicle; each of forward, backward, and downward. Because of
their narrow beam-angle, we apply simple centreline model (Thrun et al., 2005) for all of
these sonar behaviour.
Throughout its underwater operation, the vehicle is forced to keep away from the basin wall
at least 2m. In this case, the maximum range error is about 2.2m (dM in Fig. 10). Another
significant error source is misalignment of range sonar with AHRS. For vehicle's dynamics,
we observed that the yaw angular velocity of P-SURO was less than 10o/s. Consider the
1500m/s of acoustic velocity in the water, 10o/s of yaw motion may cause less than 0.1o of
azimuth error. Therefore, the effect of vehicle dynamics on the range measurement can be
neglected.
To investigate the range sonar behavior in a basin which is of a small cuboid, we performed
a simple test where the vehicle is rotated (about 7.6o/s) several cycles with the center at the
same point. Throughout the test, we force the vehicle to keep zero pitch angle. The resulted
basin profile image is shown in Fig. 11, from which we can see that closer to the basin
corner, more singular measurements are occurred.

5.2 Obstacle detection

At any point ( , , ), through rotating the vehicle on the horizontal plane, we can easily
get a 2D profile image of basin environment, see Fig. 12. And for 3D case, we simply extend
the horizontal rotating motion with constant velocity of descent/ascent motion and get
rough 3D profile image, see Fig. 13. According to this profile image, we detect the obstacle
and further design corresponding vehicle path.
The obstacle block A in Fig. 12 is modelled as ( , ) with obstacle start point and the end
point. Here = | | and is yaw angle of with = , , . In the case of the vehicle
facing point a and c, if | | = | | − | | > where = ( ) is a design parameter, then
point a is taken as the start point of an obstacle. And in the case of b and d, if | | = | | −
| |> with = ( ), then b is taken as the end point of the obstacle.

5.3 Path planning

Path planning is an important issue in the robotics as it allows a robot to get from point A to
point B. Path planning can be defined as "determination of a path that a robot must take in
order to pass over each point in an environment and path is a plan of geometric locus of the
points in a given space where the robot has to pass through" (Buniyamin et al., 2011). If the
robot environment is known, then the global path can be planned off line. And local path
planning is usually constructed online when the robot face the obstacles. There are lots of
path planning methodologies such as roadmap, probability roadmap, cell decomposition
method, potential field have been presented so far (Choset et al., 2005; Khatib, 1986;
Valavanis et al., 2000; Elfes, 1989; Amato & Wu, 1996; Li et al., 2008a).
For P-SURO AUV, there is only one range sonar mounted in front of vehicle. To get a profile
image of environment, the vehicle has to take some specific motions such as rotating around
it, and this usually takes quite an amount of time. In other word, it is not suitable for the
vehicle to frequently take obstacle detecting process. For this reason, we design a relatively
Development of a Hovering-Type Intelligent Autonomous Underwater Vehicle, P-SURO 31

simple path planning method for autonomous navigation of P-SURO AUV. Consider the
Fig. 12, to get to the point =( , , ℎ ), the vehicle will turn around at start point
. According to detected obstacle A(a, b), we calculate , and = max{ , }. If
> with design parameter, then we design the target point as
=( , , ℎ ) with = + 0.5 (see Fig. 12, in this case, we assume
> and > ). In the case of ≤ , the vehicle will take descent
motion as shown in Fig. 13 until ℎ = ℎ . Here (see Fig. 13) is a design parameter.
And in this case, the target point is set to = ( , , ℎ ).

Fig. 10. Maximum range error.

Fig. 11. Basin profile image using range sonar.

32 Autonomous Underwater Vehicles

Fig. 12. Acquisition of 2D profile image.

Fig. 13. Acquisition of 3D profile image.

6. Basin test
To demonstrate the proposed vision-based underwater localization and the SLAM methods,
we carried out a series of field tests in the engineering basin in PIRO.

6.1 Preliminary of basin test

In its underwater mission, the vehicle is always forced to keep zero pitch angle. And in the
horizontal plane, we design the vehicle's reference path to be always parallel to the axis X or
axis Y, see Fig. 12. In this case, at any point, the vehicle's position can be easily got through
simple rotation mode. However, considering the fact that the vehicle does not keep at the
Development of a Hovering-Type Intelligent Autonomous Underwater Vehicle, P-SURO 33

same point through its rotation, in other word, there is a drift for the vehicle's position in the
rotating mode. So, though the accuracy of range sonar measurement is in the centimetres
level, the total position error for this kind of rotation mode is significant. Through a number
of basin tests, we observe that this kind of position error is up to 0.5m.
Consider this kind of forward/backward motion; the vehicle's forward/backward velocity
can be calculated using range sonar measurements. For this purpose, the following filter is
designed for acquisition of range sonar raw measurements

( ) = (1 − 2 ) ( − 1) + 2 ( ), (6)
where and denote each of filtered and raw measurements of range sonar, and is
filtering order. The filtering results can be seen in Fig. 14.

Fig. 14. Calculated forward speeds.

Fig. 15. Comparison of heading measurements

34 Autonomous Underwater Vehicles

Another important issue for the basin test is about vehicle's AHRS sensor. The engineering
basin in the PIRO is located in the basement of building, which is mainly constructed by
steel materials. In this kind of environment, because of heavy distortion of earth magnetic
field, AHRS cannot make proper initialization and Kalman filter compensation process.
Therefore, there is significant drift in the AHRS heading output. However, fortunately, there
is high accuracy 1-axis Gyro sensor horizontally mounted on the vehicle for the motion
control purpose. And we estimate the vehicle's heading value using this Gyro output, whose
bias value is also evaluated through lots of basin tests. Fig. 15 shows the comparison of these
measurements.

6.2 P-SURO SLAM

To demonstrate the SLAM method proposed for P-SURO AUV, we perform the following
three autonomous navigation tests: a) without obstacle, b) with one obstacle, c) with two
obstacles, see Fig. 16. The autonomous navigation mission can be divided into following
four phases

Fig. 16. Test environment and corresponding range sonar profile images.
Obstacle Detecting Phase: At start point = ( , , , )=(3m,4m,1.5m, 90o), the vehicle turn
a half cycle counter-clock wisely. In this period, the vehicle detects the obstacle using
forward range sonar.
Path Planning Phase: According to the profile image got from the Obstacle Detecting Phase,
the vehicle designs a target point = ( , , , 0 ).
Vision-based Underwater Localizing Phase: While approaching to , the vehicle
recognizes the underwater pattern, from which defines the end point = (9, − , −
, 0 ). Here ( , , ) denotes the vehicle's current position, and ( , , ) is the vehicle's
pose information acquired from pattern recognition.
Homing Phase: After approaching , or failed to recognize the pattern, the vehicle returns
to along with its previous tracking trajectory.
Development of a Hovering-Type Intelligent Autonomous Underwater Vehicle, P-SURO 35

In Fig. 16, the blue line (calculated basin wall) is got through ( , , , , ) where
is the forward range sonar measurement.
In the Path Planning Phase, the target points are set to different values according to three
different cases. In the case of without obstacle, we set = (8 , 4 , 1.5 , 0 ); in the
case of one obstacle, set = (8 , + 0.5 , 1.5 , 0 ) with is shown in Fig.
16(b); and in the case of two obstacles, = (8 , 4 , ℎ , 0 ), where ℎ is defined in
Fig. 13 .
Autonomous navigation with obstacle avoidance and underwater pattern recognition test
results are shown in Fig. 17. Through these field tests, we found that the proposed SLAM
method for P-SURO AUV shown a satisfactory performance. Also, we found that the
aforementioned drift in the vehicle's rotating motion is the main inaccuracy source of both
the navigation and the path planning (specially, in the calculation of with = , , in
Fig. 12).

Fig. 17. Autonomous navigation test results.

7. Summary and future works

Recently, how to improve the vehicle's autonomy has been one of most hot issues in the
underwater robotics community. In this chapter, we have discussed some of underwater
intelligent technologies such as vision-based underwater localization and SLAM method
only using video camera and range sonar, both of which are relatively cheap underwater
equipments. Through a series of field tests in the engineering basin in PIRO using P-SURO
36 Autonomous Underwater Vehicles

AUV, we observed that the proposed technologies provided satisfactory accuracy for the
autonomous navigation of hovering-type AUV in the basin.
However, through the basin tests, we also observed that proposed vision algorithm was
somewhat overly sensitive to the environmental conditions. How to improve the robustness
of underwater vision is one of great interest in our future works. Besides, developing certain
low-cost underwater navigation technology with partially known environmental conditions
is also one of our future concerns.

8. Acknowledgment
This work was partly supported by the Industrial Foundation Technology Development
project (No. 10035480) of MKE in Korea, and the authors also gratefully acknowledge the
support from UTRC(Unmanned Technology Research Centre) at KAIST, originally funded
by ADD, DAPA in Korea.

9. References
Amato, N. M. & Wu, Y. (1996). A randomized roadmap method for path and manipulation
planning, Proceedings of IEEE International Conference on Robotics and Automation, pp.
113-120, Osaka, Japan, November 4-8, 1996
Bennet, A. A & Leonard, J. J. (2000). A behavior-based approach to adaptive feature
detection and following with autonomous underwater vehicles. IEEE Journal of
Oceanic Engineering, Vol.25, pp. 213-226, 2000
Brooks, R. A. (1986). A robust layered control system for a mobile robot. IEEE Journal of
Robotics and Automation,Vol.2, pp. 14-23, 1986
Buniyamin, N., Sariff, N., Wan Ngah, W. A. J., Mohamad, Z. (2011). Robot global path
planning overview and a variation of ant colony system algorithm. International
Journal of Mathematics and Computers in Simulation, Vol.5, pp. 9-16, 2011
Castellanos, J. A. & Tardos, J. D. (1999). Mobile Robot Localization and Map Building:
A Multisensor Fusion Approach. Boston, Mass.: Kluwer Academic Publishers,
1999
Choset, H., Lynch, K. M., Hutchinso, S., Kantor, G., Burgard, W., Kavraki, L. E., Thrun, S.
(2005). Principles of Robot Motion. MIT Press, 2005
Do, K. D., Jiang, J. P., Pan, J. (2004). Robust adaptive path following of underactuated ships.
Automatica, Vol.40, pp. 929-944, 2004
Elfes, A. (1989). Using occupancy grids for mobile robot perception and navigation. IEEE
Transactions on Computer, Vol.22, pp. 46-57, 1989
Khatib, O. (1986). Real-time obstacle avoidance for manipulators and mobile robots.
International Journal of Robotics Research, Vol.5, pp. 90-98, 1986
Hartley, R & Zisserman, A. (2000). Multiple View Geometry in Computer Vision. Cambridge
University Press, June 2000
Healey, A. J., Marco, D. B., McGhee, R. B. (1996). Autonomous underwater vehicle control
coordination using a tri-level hybrid software architecture, Proceedings of the IEEE
International Conferences on Robotics and Automation, Minneapolis, Minnesota, pp.
2149-2159, 1996
Development of a Hovering-Type Intelligent Autonomous Underwater Vehicle, P-SURO 37

Jiang, J. P. (2002). Global tracking control of underactuated ships by Lyapunov's direct

method. Automatica, Vol.40, pp. 2249-2254, 2002
Marthiniussen, R., Vestgard, K., Klepaker, R., Storkersen, N. (2004). HUGIN-AUV concept
and operational experience to date, Proceedings of IEEE/MTS Oceans'04, pp. 846-850,
Kobe, Japan, November 9-12, 2004
Leonard, J. & Durrant-Whyte, H. (1992). Directed Sonar Sensing for Mobile Robot Navigation.
London: Kluwer Academic Publishers, 1992
Li, J. H., Jun, B. H., Lee, P. M., Hong, S. K. (2005). A hierarchical real-time control
architecture for a semi-autonomous underwater vehicle. Ocean Engineering, Vol.32,
pp. 1631-1641, 2005
Li, J. H., Lee, P. M., Jun, B. H., Lim, Y. K. (2008a). Underwater Vehicle, InTech, ISBN 978-953-
7619-49-7, Vienna, Austria
Li, J. H., Lee, P. M., Jun, B. H., Lim, Y. K. (2008b). Point-to-point navigation of underactuated
ships. Automatica, Vol. 44, pp. 3201-3205, 2008
Li, J. H., Yoon, B. H., Oh, S. S., Cho, J. S., Kim, J. G., Lee, M. J., Lee, J. W. (2010). Development
of an Intelligent Autonomous Underwater Vehicle, P-SURO, Proceedings of
Oceans'10 IEEE Sydney, Sydney, Australia, May 24-27, 2010
Oh, S. S., Yoon, B. H., Li, J. H. (2010). Vision-based localization for an intelligent AUV,
P-SURO, Proceedings of KAOSTS Annual conference, pp. 2602-2605, Jeju, Korea,
2010
Peuch, A., Coste, M. E., Baticle, D., Perrier, M., Rigaud, V., Simon, D. (1994). And advanced
control architecture for underwater vehicles, Proceedings of Oceans'94, Brest, France,
pp. I-590-595, 1994
Prestero, T. (2001). Verification of a six-degree of freedom simulation model for the REMUS
autonomous underwater vehicles, Masters Thesis, MIT, USA
Quek, C & Wahab, A. (2000). Real-time integrated process supervision. Engineering
Applications of Artificial Intelligence, Vol.13, pp. 645-658, 2000
Simon, D., Espiau, B., Castillo, E., Kapellos, K. (1993). Computer-aided design of a generic
robot controller handling reactivity and real-time control issues. IEEE Transactions
on Control Systems Technology, Vol.1, pp. 213-229, 1993
Thrun, S., Burgard, W., Fox, D. (2005). Probabilistic Robotics. The MIT Press, 2005
Valavanis, K. P., Hebert, T., Kolluru, R., Tsourveloudis, N. C. (2000). Mobile robot
navigation in 2-D dynamic environments using electrostatic potential fields.
IEEE Transactions on Systems, Man and Cybernetics-part A, Vol.30, pp. 187-197,
2000
Wang, H. H., Marks, R. L., Rock, S. M., Lee, M. J. (1993). Task-based control architecture for
an untethered, unmanned submersible, Proceedings of the 8th Symposium on
Unmanned Untethered Submersible Technology, Durham, New Hampshire, pp. 137-
148, 1993
Zhang, Z. (2000). A flexible new technique for camera calibration. IEEE Transactions
on Pattern Analysis and Machine Intelligence, Vol. 22, No. 11, pp. 1330-1334,
2000
38 Autonomous Underwater Vehicles

Zheng, X. (1992). Layered control of a practical AUV, Proceedings of the Symposium

on Autonomous Underwater Vehicle Technology, Washington DC, pp. 142-147,
1992
 3

Hydrodynamic Characteristics of the Main Parts

of a Hybrid-Driven Underwater Glider PETREL
Wu Jianguo1, Zhang Minge2 and Sun Xiujun2
1Shenyang Institute of Automation Chinese Academy of Sciences
2Tianjin University

China

1. Introduction
Autonomous Underwater Vehicle (AUV), Remotely Operated Vehicle (ROV) and
Autonomous Underwater glider (AUG) are the main autonomous underwater platforms
available currently, which play important role in the marine environmental monitering. The
relationships between those three types of vehicles were shown in Figure 1.

Fig. 1. Underwater Vehicles

As a special type of AUV, underwater gliders have many advantages, such as long
endurance, low noise and low energy cost. A glider can periodically change its net
buoyancy by a hydraulic pump, and utilize the lift from its wings to generate forward
motion. The inherent characteristics of a glider can be summarized as buoyancy-driven
propulsion, sawtooth pathway, high endurance and slow speed. There exist three legacy
gliders named respectively Seaglider, Spray and Slocum [1~6]. In spite that underwater
gliders features low level of self noise and high endurance, they also have weaknesses like
the lack of maneuverability and the inability to perform a fixed depth or level flight [7].
Driven by a propeller with carried energy source, autonomous underwater vehicles is
preprogrammed to carry out an underwater mission without assistance from an operator on
the surface. However, they can only cover a relatively short range after each recharge due to
the high power consumed for propulsion and generate much more noise than the AUGs
because of its propeller and motors [8~10]. The range of AUV’s is restricted by the amount of
energy carried on board, can was not more than several hundreds kilometers in general [11].
The performances of the underwater vehicle are compared in Figure 2.
40 Autonomous Underwater Vehicles

Low Time (range)

No limitation（1km）
Endurance

ROV High

Maneuverability
Several hours（100km）
AUV

Several months（1000km）
High AUG Low

Fig. 2. Performances of three Underwater Vehicles

By combining the advantages of the glider and the propeller-driven AUVs, A hybrid-driven
underwater glider PETREL with both buoyancy-driven and propeller-driven systems is
developed. Operated in buoyancy-driven mode, the PETREL carries out its mission to
collect data in a wide area like a legacy glider. When more exact measurements of a smaller
area or level flight are needed, the PETREL will be operated by using the propeller-driven
system [5, 7]. This flexible driven glider contributes to have a long range while operated in the
buoyancy driven mode like a glider, as well as improve the robust performance to deal with
some wicked circumstances by the propeller driven system [7].
Proper hydrodynamic design is important for the improvement of the performance of an
underwater vehicle. A bad shape can cause excessive drag, noise, and instability even at low
speed. At the initial stage of design, there are two ways to obtain the hydrodynamic data of
the underwater vehicle, one is to make model experiment and the other is to use the
computational fluid dynamics (CFD). With the development of the computer technology,
some accurate simulation analysis of hydrodynamic coefficients have been implemented by
using the computational fluid dynamic (CFD) software, instead of by experiments at a much
higher cost over the past few years [12-13] . In consideration of the reduced time, lower cost,
more flexible and easier optimumal design, the CFD method was used in this article. The
fluent Inc.’s (Lebanon,New Hampshire) CFD software FLUENT 6.2 was adopted by this
article.
This chapter focuses on the hydrodynamic effects of the main parts of a hybrid-driven
underwater glider especially in the glide mode. By analyzing the results of the three main
hydrodynamic parts, the wings, the rudders and the propeller, the characteristics of drag,
glide efficiency and stability will be discussed, and suggestions for altering the HUG’s
design to improve its hydrodynamic performance are proposed.

2. Computational details
2.1 Mathematical model
A criterion for determining of the flow regime of the water when the vehicles moving in it is
proposed by Reynolds number [14-15]:

R e = ρ vL μ
(1)
Here ρ is the density of water, v is the velocity of vehicle, L is the characteristic length,
μ is the dynamic coefficient of viscosity. The transition point occurred when the Reynolds
Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 41

number is near 106 for the external flow field, which is called critical Reynolds numbers. It
was laminar boundary layer when the Re < 5 × 10 5 , it was seem as turbulent flow while
Re > 2 × 106 . The Reynolds number of the hybrid underwater glider PETREL at two
different steering modes is shown in table 1.

steering mode velocity v / (m/s) Reynolds number

Glider 0.5 1.25×106
AUV 2 5×106
Table 1. The Reynolds number at different steering modes
The turbulence model will be adopted because the Reynolds numbers of the PETREL in two
steering modes are all above the critical Reynolds numbers. Computations of drag, lift and
moment and flow field are performed for both the model over a range of angles of attack by
using the commercially available CFD solver FLUENT6.2. The Reynolds averaged Navier–
Stokes equation based on SIMPLAC algorithm and the finite volume method were used by
our study. In our study RNG k-ε model was adopted and the second-order modified scheme
was applied to discrete the control equations to algebra equations. Assuming that the fluids
were continuous and incompressible Newtonian fluids. For the incompressible fluid, the
RNG k − ε transport equations are [12, 16]:

∂ ∂ ∂ ⎛ ∂k ⎞
ρ (k) + ρ ( kui ) = ⎜ α k μ eff ⎟ + Gk − ρε + Sk (2)
∂t ∂xi ∂x j ⎜⎝ ∂x j ⎟⎠

∂ ∂ ∂ ⎛ ∂ε ⎞ ε ε2
ρ (ε ) + ρ (ε ui ) = ⎜ α ε μ eff ⎟ + C 1ε Gk − C 2 ε ρ − Rε + Sε
∂t ∂xi ∂x j ⎜⎝ ∂x j ⎟⎠ k k (3)

Here Sk and Sε are source items, μ eff is effective viscosity, Gk is turbulence kinetic energy
induced by mean velocity gradient.

∂u j
Gk = − ρ ui' u'j
∂xi (4)

σ k and σ ε is respectively the reversible effect Prandtl number for k and ε .

C 1ε = 1.42 , C 2ε = 1.68
In the RNG model,a turbulence viscosity differential equation was generated in the non-
dimensional treatment.

⎛ ρ 2k ⎞ νˆ
d⎜ ⎟⎟ = 1.72 3 dνˆ
⎜ εμ ν − 1 + Cν
ˆ (5)
⎝ ⎠

μ eff
here, νˆ =
μ , Cν ≈ 100 . Taking the integral of the(5),the exact description of active
turbulence transport variation with the effective Reynolds number can be acquired, which
makes the mode having a better ability to deal with low Reynola number and flow near the
wall. For the large Reynola number, the equation(5)can be changed into (3-6).
42 Autonomous Underwater Vehicles

k2
μ t = ρC μ (6)
ε
Here , C μ = 0.0845 The RNG k − ε model was adopted due to the initial smaller Reynola
number of boundary layer, and the more exact results can be gained by substituting the
differential model into the RNG k − ε model.

2.2 Meshing and boundary conditions

The size function and unstructured meshes were adopted to keep the meshes distributing
reasonably and make the meshes generating expediently. The examples of meshing are
shown as figure 3, figure 4 and figure 5.

Fig. 3. Stern meshes

Fig. 4. Two-dimension rudder meshes

Fig. 5. Whole meshes of the vehicle

Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 43

Boundary conditions:
1. inlet boundary condition: setting the velocity inlet in front of the head section with a
distance of one and a half times of the length .
2. outlet boundary condition: setting the free outflet behind the foot section with a
distance of double length of the vehicle.
3. wall boundary condition: setting the vehicle surface as static non-slip wall.
4. pool wall boundary condition: non-slip wall.

2.3 Results verification

To verify the precision of the calculation, we computed the drag coefficients of Slocum
underwater glider [17] at different angle of attack as shown in table 2. The table 3 shows the
verification of numerical simulation results of drag of AUV shell of Tianjin University. The
error percentage of our calculation is less than 9.35％.

Angle of attack Reynolds number CD CD Error

α(degree) Re (experiment) (CFD) percentage

-2.9 7.5×105 0.31 0.281 9.35%

2.3 6.3×105 0.25 0.268 7.20%

2.7 5.8×105 0.27 0.274 1.46%

Table 2. Verification of numerical simulation results of CD

Velocity Reynolds Drag Error

Drag Experiment(N)
/(m/s) number CFD(N) percentage

0.81 2.5×106 7.4 6.903 6.72%

1.4 4.4×106 20.3 19.92 1.87%

2.0 6.2×106 37.5 37.34 0.427%

Table 3. Verification of numerical simulation results of drag of AUV shell

3. Wing hydrodynamic design [18]

3.1 Orthogonal experimental design and results analysis
3.1.1 Orthogonal experimental design
An orthogonal experimental with four factors and three levels was conducted by keeping
the main body size of the vehicle as constant. The four factors are wing chord, aspect ratio,
backswept and distance between the center of wing root and the center of body. The
simulation experiments were done at the situation of angle of attack is α = 6° and the
velocity is v = 0.5m / s . The airfoil of the wings was NACA0010. The orthogonal
experimental table was shown as table 4.
44 Autonomous Underwater Vehicles

level chord(mm) aspect ratio backswept (°) distance(mm)

1 100 6 20 100
2 150 8 40 0
3 200 10 60 -100

Table 4. Orthogonal experimental table

3.1.2 Analysis indexes

The design of the wing will generate important impacts on glide efficiency and glide
stability of the vehicle. The lift to drag ratio L D is chosen for measurement of the glide
efficiency, the bigger values correspond to the more efficient gliding. The inverse of L/D
expresses the glide slope [7, 19]. Existing oceanographic gliders are designed for static stability
in steady glides, and the static stability can be measured by the non-dimensional
hydrodynamic lever lα' , the equations are[20~21]:

lα' = lα / l
(7)

lα = − Mα / Lα
(8)
Here, l is the vehicle length, Mα is the hydrodynamic moment induced by angle of
attack α , Lα and is the Lift induced by the angle of attack α . It is static instability
while lα' > 0 , the moment induced by incremental angle of attack makes the angle of attack
become bigger; It is neutral stability while lα' = 0 ; It is called static stability while lα' < 0 , the
moment induced by incremental angle of attack makes the vehicle to turn to the original
state.

3.1.3 Influencing factors analysis

The orthogonal experimental table L18(37) and the simulation results are shown in table5.
The trend charts were shown as Figure 6 and Figure 7. The L / D increase with the growth
of chord and aspect ratio, and decrease with the growth of backswept,it has little
relationship with the location of the wings. The lα′ increase as chord and backswept increase
when the wings is located after the hydrodynamic center, which means the stability increase
as chord and backswept raise. The stability gets higher as the wing location becomes father
away from behind the center of the body.
The ranges of chord, aspect ratio, backswept and distance of the wings is separately 2.448,
1.077, 1.303 and 0.312 for the L/D, which was gained by the range method. It is shows that
the effects significance series for glide efficiency is chord, backswept, aspect ratio and the
location of wings. The chord was dramatic for the index L/D at the significance level 0.10
and 0.05 adopted by the range method.
In like manner, the range of chord, aspect ratio, backswept and distance of the wings is
separately 0.051, 0.037, 0.095 and 0.031 for the lα′ , which was gained by the range method.
It is shows that the effects significance series for glide stability is backswept, chord, aspect
ratio and the location of wings. The backswept was dramatic for the index lα′ at the
significance level 0.10 and 0.05 adopted by the range method.
Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 45

Simulation factors results

times chord(mm) aspect ratio backswept (°) distance(mm) L /D lα′

1 100 6 20 100 2.86 0.0274

2 100 8 40 0 3.11 -0.0323
3 100 10 60 -100 2.74 -0.109
4 150 6 20 0 4.07 -0.00059
5 150 8 40 -100 4.37 -0.0854
6 150 10 60 100 3.82 -0.127
7 200 6 40 100 4.88 -0.0158
8 200 8 60 0 4.47 -0.172
9 200 10 20 -100 6.81 -0.0713
10 100 6 60 -100 2.33 -0.0630
11 100 8 20 100 3.31 0.0212
12 100 10 40 0 3.44 -0.0444
13 150 6 40 -100 3.78 -0.0594
14 150 8 60 100 3.40 -0.0858
15 150 10 20 0 5.30 -0.0237
16 200 6 60 0 3.94 -0.115
17 200 8 20 -100 6.17 -0.0562
18 200 10 40 100 6.21 -0.0756

Table 5. Orthogonal experimental table and the results

chord (mm) aspect ratio backswept location (mm)

Fig. 6. L/D tendency chart
46 Autonomous Underwater Vehicles

chord (mm) aspect ratio backswept location (mm)

Fig. 7. lα′ Tendency chart

It is well known that the chord and aspect ratio of the wings should be increased, and the
backswept decreased for the higher glide efficiency when PETREL is operated in the gilde
mode. Simutaneously, the backswept of the wings should be increased and the wings
should be moved backward father behind the center of the vehicle for the higher stability. It
indicates that the effects from the increment of the backswept of the wings are inversed in
increasing the glide efficiency and the stability. The backswept of the wings should be
determined in terms of other capability indexes of the underwater vehicle.

3.2 Concrete models analysis

Four concrete models with varied wing parameters listed in table 6 were choosen for some
further investigation. We carried out this new series of experiments in the hope of providing
the effects of wings, rudders and propeller on the L / D and lα′ at different glide angle of
attack when the velocity is 0.5m/s and the angle of attack on the rang of 0°~20°.The model
one has the highest glide efficiency and glide stability in the table6; The models 2~4 were
proposed in order to evaluate the affects of location, aspect ratio and chord of wings on the
analysis index as shown in the table 6. The Figure 8 gives the pressure distribution chart of
model 3. The calculation results of different models are shown as Fig. 9~ Fig. 12.

models chord(mm) aspect ratio backswept(°) location(mm)

1 200 10 40 0
2 200 10 40 100
3 200 8 40 0
4 150 10 40 0
Table 6. The parameter of the concrete model
The location of the wings has little influence on the L/D, which means it has little influence
on the glide efficiency illustrated in Figure 9, but it has dramatic effects on the glide stability
which can be seen in the Figure 10. From the figure 9 and 10, it can be seen that the L/D
decreased and the lα′ increased obviously when the aspect ratio and chord reduced, but the
effects is more dramatically to decrease the chord of the wings for the L/D. It has the biggest
lift to drag ratio when the angle of attack at 6 degree shown in Figure 9, that is means the
maximum glide efficiency can be gain when the angle of attack at 6 degree.
Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 47

2
1.33×10
1
7.82×10

2.35×10
1
-1.75×10
1
-5.85×10
1
-9.96×10
2
1 41×10

Fig. 8. The pressure distribution chart of model 3

7
model 1
6
model 2
Lift-to-Drag Ratio L/D

5 model 3
4 model 4

3
2
1
0
0 2 4 6 8 10 12 14 16 18 20
angle of attack α(°)

Fig. 9. The relationship between L/D and angle of attack

0
model 1
-0.02
lα’

model 2
-0.04 model 3
Static Stability Coefficient

model 4
-0.06
-0.08
-0.1
-0.12
-0.14
-0.16
2 4 6 8 10 12 14 16 18 20
angle of at tack α(°)

Fig. 10. The relationship between lα′ and angle of attack

48 Autonomous Underwater Vehicles

The drag ratio of rudders and propeller

0.4
model 1
0.35
model 2
0.3 model 3
0.25 model 4

0.2
0.15
0.1
0.05
0
0 2 4 6 8 10 12 14 16 18 20
angle of at t ack α(°)

Fig. 11. The drag ratio of rudders and propeller

lift 6
coefficient
5

4
Lift Coefficient C L

3 model 1
model 2
2 model 3
model 4
1

-1
0 0.5 1 1.5 2
drag coefficient C D

Fig. 12. The lift-drag polar curve of four concrete models

The drag of the hybrid glider will be increased because of the drag generated by the rudders
and propeller compared with the legacy gliders in the glide mode. The range in the glide
mode will be decreased because of the drag of these parts . The ratio of drag on the propeller
and rudders to whole drag is illustrated in Fig. 11, where we find that the ratio changed as
the angle of attack increases, and the values is within the range of 10％~35％. Compared
with the legacy gliders, the range of the vehicles with the same configuration as PETREL
will be decrease 10％~35％. The Lift to Drag polar curves of the four concrete models are
shown as figure 12. The model 3 and model 4 have the bigger lift than the model 1 and
model 2 when the drag coefficients from the figure 3-9 is less than 0.5, but the lift of model 1
and model 2 increases greatly when drag coefficients gets bigger than 0.5. Due to the drag of
the vehicle need overcome by the variable buoyancy B in the end and there is equation (9),
so the net buoyancy supplied by the buoyancy driven system and glide angle should be
taken into consideration.

B sin θ = D (9)
Here B is the net buoyancy, θ is the glide angle, D is the drag of the glider.
Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 49

3.3 Results and discussion

The orthogonal experiment shows that glide efficiency is most significantly influenced by the
chord length while stability of the vehicle is most remarkably affected by the sweep angle.
Further numerical calculations based on four specific models with the attack angle in the
range of 0°-20° indicate that location of the wings mainly affects glide stability but has little
influence on glide efficiency.
When the vehicle glides at about 6° attack angle it has the maximum ratio of lift to drag. The
range of the hybrid glider with the same configuration as PETREL will be decrease
10％~35％ compared with the legacy gliders.

4. Rudder hydrodynamic design [22]

4.1 Rudder parameters
The rudders parameters include root chord, half span, aerofoil and backswept, which are
shown in Figure 13. As defined in the [23], the chord is denoted by C , the distance from the
leading edge to trailing edge in a given two-dimensional section. The chord is measured in
parallel with the section at the root of the rudder. In general, the chord can vary along the
span, in which case the geometric mean chord, C , is used in computations unless noted[21].
The C is defined based on Figure 14 as

Ct + Cr
C= (10)
2

foil section
Tip chord Ct
thickness
leading edge

trailing edge

semi-span b/2

Root Chord Cr

Fig. 13. Rudders parameters

Rudder post

Fig. 14. Foil section and hydrodynamic force

50 Autonomous Underwater Vehicles

The semi-span, denoted by b / 2 , measures the distance from the rudder root to tip along the
line perpendicular to the root section. The span, in this work, is twice as long as the root-to-
tip distance for an isolated plan. The hydrodynamic forces including lift and drag acted on
the aerofoil is shown in Figure 14 and can be expressed as

L = 1 2 ρC L AV 2
(11)

D = 1 2 ρC D AV 2
(12)
Here, ρ is the density of the water; C L is the lift coefficient; C D is the drag coefficient; A is
the area of rudder; V is the velocity of water; α is the angle of attack. The rudderpost
location is expressed by P , which is shown in Figure 14.

4.2 Foil section

The geometry of a rudder is mainly defined by the two-dimensional foil section. The
symmetrical foil sections are generally used by the underwater vehicles. Many types of the
foil sections are proposed by many countries to improve the hydrodynamic performance.
The famous foil sections series include NACA series, HEЖ series, ЦАГИ series, and JFS
series [21], among which the four-digit NACA sections are most widely used for underwater
vehicle rudders in that it provides the higher lift and the lower drag. The four-digit NACA
section series is a low velocity foil sections series, and have a bigger radius of leading edge
and a plumpy head section, which is suitable for the rudder of underwater vehicles at low
velocity. In this work, the four digit NACA00×× section was used, where the ×× denote the
thickness-to-chord ratio. The lift coefficient and drag coefficient of the foil sections can be
calculated as

L
CL =
1 2 ρV 2C
(13)

D
CD =
1 2 ρV 2C
(14)
Here, L is the profile lift, D is the profile drag, C is the chord. The NACA0008, NACA0012,
NACA0016, NACA0020 and NACA0025 are usually used for the rudders of miniature
underwater vehicles, their hydrodynamic characteristics were calculated by using
computational fluid dynamics. According to the most often adopted velocity of the
autonomous underwater vehicles and the velocity of PETREL in AUV mode, the calculation
velocity was determined as 2m/s. An example of CFD meshing result is shown in figure 15,
where the unstructured mesh was adopted and the wall of section was made dense. The
calculating results were shown in the Figure 16~ Figure 18
The relationship of lift coefficient and angle of attack is illustrated in Figure 16, where we
can see that there was a bigger angle of stalling and bigger maximal lift coefficient when the
section becomes much thicker. From the figure 17 we can see that the thinner wing section
has a lower drag cofficient when the angle of attack is small, but the thicker wing section has
a lower drag cofficient when the angle of attack is bigger than a certain critical angle of
attack. The NACA0008 section has the maximal L/D and NACA0025 has the minimal L/D
Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 51

than other sections which is shown in the Figure 18. The NACA0012 section with angle of
stall about 20° and a higher L/D was adopted by the Hybrid glider PETREL.

Fig. 15. CFD meshing results

1.25
1
0.75
CL

0.5 NACA0008
NACA0012
0.25
Lift Coefficient

NACA0016
NACA0020
0
NACA0025
-0.25
-0.5
-0.75
-1
-1.25
-40 -30 -20 -10 0 10 20 30 40
angle of at t ack α(°)

Fig. 16. The relationship of profile lift coefficient and angle of attack

0.7
NACA0008
0.6 NACA0012
NACA0016
NACA0020
Drag Coefficient C D

0.5 NACA0025

0.4

0.3

0.2

0.1

0
-40 -30 -20 -10 0 10 20 30 40
angle of attack α(°)

Fig. 17. The relationship of profile drag coefficient and angle of attack
52 Autonomous Underwater Vehicles

10
8
6
4
2
L/D 0
-2
NACA0008
-4 NACA0012
NACA0016
-6 NACA0020
NACA0025
-8
-10
-40 -30 -20 -10 0 10 20 30 40
angle of at tack α(°)

Fig. 18. The relationship of L/D and angle of attack

4.3 Area of rudder calculation

The area of rudder as an important parameter for maneuverability of the underwater
vehicle is related to the size and shape of the body. The area of rudder can be design by cut
and try method, master model method and empirical formula design method. For the high
maneuverable ship, the control surfaces can be designed according to Det Norske Veritas,
(DNV) rudder sizing rules [24].

DL B
Area = [1 + 25( )2 ]
100 L (15)
Here, D is the diameter of the vehicle, L is the length of the vehicle, B is the width of the
vehicle, and B = D for revolution body. It suggested 30% increase in area if rudders in front
of the propeller, and then increased by an additional 50% to match empirical data from
other underwater vehicles by the DNV rules. The turn diameter induced by single rudder is
about triple-length of the vehicle in terms of the design by DNV rules. The rudder design for
the hybrid glider PETREL is shown in Figure 19 and the parameters of the rudder shown in
table 7.

Fig. 19. The photo of the rudder of PETREL

Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 53

parameters Tip chord Ct Root chord C r Semi-span b / 2 section

Value 125mm 200mm 120mm NACA0012
Table 7. The parameters of the rudder

4.4 Hinge moment analysis

The hinge moment is produced by a hydrodynamic force about the hinge line of a control
surface. It makes an impact on maneuverability of the underwater vehicle in that the hinge
moment must be overcome during steering. The bigger hinge moment will make the turning
velocity of rudders become slowly and make the control action slow-witted.
The hydrodynamic performance of three dimension rudders at different angles of attack
was simulated by using CFD methods. The inlet velocity was set to be 2m/s.

2.6×103
1.4×103
presssure P/Pa

2.6×102
-5.2×102
-1.7×103
-2.5×103
-4.1×103
-5.2×103

Fig. 20. Pressure distribution chart when angle of attack is 20°

1.2

CL
1
CD
0.8 L/10D
CL,CD,L/10D

0.6

0.4

0.2

0
0 4 8 12 16 20 24 28 32 36 40
angle of attack α
(°)

Fig. 21. C L , C D and L / 10D variation curve with different angles of attack
54 Autonomous Underwater Vehicles

0.3
0.25 P=0.2c

Hinge Moment Coefficient C M

P=0.25c
0.2 P=0.3c
P=0.35c
0.15 P=0.4c
P=0.45c
0.1 P=0.5c
P=0.55c
0.05
0
-0.05
-0.1
-0.15
0 4 8 12 16 20 24 28
angle of attack α(°)

Fig. 22. Hinge moment with different angles of attack

The pressure distribution of the rudder is illustrated in Figure 20, where we can find that
there is higher pressure on the front flow face and was local higher pressure area on the
back flow face of the tail, that means there exist roundabout flow at the tail of the rudder.
Figure 21 shows the relationship between lift, drag and angle of attack. The relationship
between L/D and angle of attack is also illustrated in the figure 21, the L/D value reduces ten
times for the same scale with other two curves. It can be known that the maximal lift to drag
ratio was about 8° and the angle of stall about 34°, so the angles of stall of three dimensional
rudders are greater than two-dimension section. The hinge moment of rudders with
different axis of rudder position is shown in Figure 22, where we can seen that the hinge
moment varied with the angle of attack. The hinge moments are little while P = 0.4c for the
rudder we design no matter how the angle of attack changed.

4.5 Results and discussion

Aiming at the key problems of the rudder design for autonomous underwater vehicle,the
hydrodynamic characteristic of the NACA00xx series section at different angles of attack
were simulated when velocity was 2m/s by using the two-dimensional computational fluid
dynamics (CFD). For the rudder we design, the stall angle is about 34° for the three
dimensional rudders and about 20° for the two-dimensional foil section, so the angle of stall
of three dimensional rudders are greater than two-dimension foil section. The area of the
rudder of PETREL was calculated using the DNV rules；The hinge moments are little
when P = 0.4c for the rudder we design no matter how the angle of attack changed.

5. Shroud hydrodynamic effects analysis[25]

For the PETREL, the propeller plays a significant role in the vehicle’s hydrodynamic
performance, so analysis of the hydrodynamic effect of a propeller with a shroud on a winged
HUG was performed with Fluent Inc.’s (Lebanon,New Hampshire) CFD software FLUENT6.2.

5.1 Models description

To analyze the effects of the shroud, two simulations were performed, where one model is
with the shroud and the other without. Two models are shown in Figure 23.
Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 55

(a) model 1 (b) model 2

Fig. 23. The models studied in the paper

5.2 Effect of shroud on the glide drag

The drag on the vehicle can be expresses as equation (16).

1
D= ρV 2C D A (16)
2

Where, D is the force of drag in Newton, ρ is the density of water in kg/m³, V is the
velocity of the vehicle in m/s, A is the reference area in m², C D is the drag coefficient
(dimensionless). The reference area A of the PETREL is 0.096m2.
Figure 24 shows the overall drag of the two models in the glide mode. The propeller in this
mode doesn’t rotate. The overall drags of two models are calculated by CFD firstly and then
are fitted by the semi-empirical formulae (16). The drag coefficients of two models are
respectively 0.32 and 0.26. The average relative error of overall drag between CFD and semi-
empirical formulae is 4.7%. The overall drag increase 21%-26% with the propeller shroud
compared with the model two according to the CFD computation results, so the shroud
greatly increased the drag of the hybrid in glide mode. The drag components of the mode1
at the speed of 0.5m/s without angle of attack are shown in Fig. 25. The drag on the body,
rudders and wings is mainly viscous forces, while the drags on the propeller, shroud and
GPS antenna pole are primarily the pressure forces,. As shown in Figure 26, the propeller
and its shroud make up over 30% of total resistance and the percentage will increase with
the increment of the velocity. The reason for the high percentage is because of the great
pressure drags on the shroud in the glide mode. The local velocity streamline diagram near
the shroud of model one shown in the Figure 27. In the Figure, we can see that
vin and Pin are the velocity and pressure inside the shroud of water, vout and Pout are the
velocity and pressure outside the shroud of water. Because the propeller doesn’t rotate in
the glide mode, the velocity of water inside the shroud is slower than that outside the
shroud, so there exits vout > vin . According to the Bernoulli equation there was Pin > Pout , so
a pressure force f is produced by the pressure difference. The percentage of the shroud
drag to total resistance is 26%-35% at the different speed due to the pressure force in the
glide mode.
56 Autonomous Underwater Vehicles

70
model1: CFD
60 model1: empirical formula

50 model2: CFD

force (N)
model2: empirical formula
40
30
20
10
0
0 0.5 1 1.5 2 2.5
velocity (m/s)

Fig. 24. The overall drags of the two models at difference velocities

2
presure drag
1.5 viscous drag
force (N)

0.5

0
body gps propeller shourd rudders wings

Fig. 25. The drag components of the mode 1 at the speed 0.5m/s
percentage of drag to total

60 m ode1;body
m odel1:GPS
50
m odel1:propeller
resistance %

m odel1:shroud
40
m odel1:rudder
30 m odel1:wings
m odel2:body
20 m odel2:gps
m odel2:propeller
10 m odel2:rudder
m odel2:wings
0
0 0.5 1 1.5 2 2.5
velocity m/s

Fig. 26. The drag distribution of vehicle at the different velocities

Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 57

vout , Pout f1
5.6×10-1
5.0×10-1
4.5×10-1 vin , Pin f1
3.9×10-1
Velocity V（m/s） f
3.7×10-1
3.4×10-1
f1
2.2×10-1
1.7×10-1 vin , Pin
1.1×10-1
5.6×10-2
0.0×10-2
vout , Pout f1

Fig. 27. The local velocity streamline diagram of model1 ( v = 0.5m / s )

5.3 Effect of shroud on the glide efficiency

The specific energy consumption can be defined using classical aerodynamics [7] as

DU Bw w D C D
Ee = = = = = (17)
Bu Bu u L C L

Underwater gliders will have a higher glide efficiency when Ee is lower. So the lift to drag
ratio L/D is a measure of glide efficiency, where bigger values represent higher glide
efficiency [7].
The Lift-to-drag ratio versus angle of attack is plotted in Fig. 28, the relations of model one is
indicated by the solid lines. The Lift-to-drag ratio of model one is lower than the model two at
different angles of attack, that means the vehicle with the shroud will have a lower glide
efficient than that without. The Lift-to-drag ratio of model one is less than model two by 20%
to 5% for the varied angles of attack within the range from 2°to 20°. The maximum lift-to-drag
ratio occurred at the angle of attack 6°-8°for both the models at different speed.

Fig. 28. Lift-to-drag ratio versus angle of attack

58 Autonomous Underwater Vehicles

5.4 Effect of shroud on the glide stability

The underwater gliders usually are designed for static stability [17], the dimensionless
hydrodynamic moment arm lα' often used to represent the static stability of the underwater
vehicles motion. The equations of the lα' are shown in equations(7)and (8).
Existing oceanographic gliders are designed to be static stable in steady glides for the easy
control and high energy economy. The hybrid-driven underwater glider PETREL was
designed as static stability for the high energy economy in the glide mode.

-0.06

-0.05

-0.04
l a'

-0.03 0.5m /s
1.0m /s
-0.02 1.5m /s
2.0m /s
-0.01
0
0 2 4 6 8 10 12 14 16 18 20 22

angle of attack α(°)

Fig. 29. The static stability coefficient lα' versus angle of attack of model one

-0.06
-0.05
-0.04
-0.03
la'

0.5m/s
-0.02 1.0m/s
1.5m/s
-0.01 2.0m/s

0
0 2 4 6 8 10 12 14 16 18 20 22
angle of attack α(°)

Fig. 30. The static stability coefficient lα' versus angle of attack of model two

Figure 29 show the static stability coefficient lα' versus angle of attack of model one and
model two. It is static stability for both of the two models in terms of our design intention.
The stability decreases when the angle of attack gets bigger than 8°, but the stability slightly
increases for model one when the angle of attack is more than 12°. The glide speed has little
effect on the stability as shown in the Figure 29 and Figure 30. Figure 31 shows the moment
of the shroud versus angle of attack of model one. The values of the moment were positive
when the angle of attack is lower than 8° for the v = 0.5 m/s and v = 1 m/s, and the angle of
attack is less than 10° for the v = 1.5 m/s and v = 2.0 m/s. The values of the moment were
negative when the angle of attack gets higher than those critical angles. So the effect of
Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 59

shroud on the static stability of model one is that, when the angle of attack is lower than the
critical angle the shroud will makes the stability decreasing but makes the stability
increasing when the angle of attack is higher than the critical angle. as shown in the Figure
31, the action of the shroud makes the stability slightly increased when the attack angle is
higher than 12° .

4
2
moment of the shroud

0
-2
-4
-6 0.5m/s
1.0m/s
1.5m/s
-8 2.0m/s

-10
0 2 4 6 8 10 12 14 16 18 20 22
angle of attack α(°)

Fig. 31. The moment of the shroud versus angle of attack of model one

5.5 Conclusions
It was found that overall drag increased by 21 to 26 percent for the model with the propeller
shroud compared with the one without a shroud, but with the same structure and size, the
shroud’s resistance is mainly pressure force.
The shroud made the lift-to-drag of the vehicle in glide mode decrease by as much as 20
percent when the angle of attack was 2º. As the angle of attack increased, the shroud’s effect
was minimized, and the decrease in lift-to-drag ratio ranged down to five percent at an
angle of attack of 20º, meaning glide efficiency decreased due to the propeller shroud.
Finally, the shroud decreases the stability of the HUG when the angle of attack is lower than
the critical angle, but increases it when the angle of attack is higher than the critical angle.
The critical angle is between 8º and 10º for velocities lower than one meter per second, and
between 10ºand 12ºfor velocities in the range of one to two meters per second.
These findings indicate that for an underwater glider, the shroud will increase drag and
decrease the glide efficiency, but it is good for stability when the angle of attack is larger
than 8º. Therefore, the shroud is not a successful design element for the HUG in glide mode,
but in propeller mode the shroud can increase the thrust of the vehicle.
Using CFD to analyze the shroud’s hydrodynamic effects shows that the vehicle should
only be equipped with this feature for activities requiring operation in propeller mode.

6. Flow field analysis

6.1 Velocity field
The direct route flow field with the velocity of the hybrid underwater glider PETREL at
0.5m/s、1m/s、1.5m/s and 2m/s was simulated by using CFD ways. The simulation
results are shown in Figure 32.
60 Autonomous Underwater Vehicles

Unit：m/s
Unit：m/s 5.6×10-1
1.1×1.0
5.0×10-1
1.0×10
4.5×10-1
9.1×10-1
3.9×10-1
7.9×10-1
3.7×10-1
6.8×10-1
3.4×10-1
5.7×10-1
2.2×10-1
4.5×10-1
1.7×10-1
3.4×10-1
1.1×10-1
2.3×10-1
5.6×10-2
1.1×10-12
2
0.0×10-2
(a) V =0.5m/s (b) V =1m/s

Unit：m/s
Unit：m/s
1.7×10-0
2.3×10-0
1.5×10-0
2.1×10-0
1.4×10-0
1.8×10-0
1.2×10-0
1.6×10-0
1.0×10-0
1.4×10-0
8.5×10-1
1.1×10-0
6.8×10-1
9.1×10-1
5.1×10-1
6.8×10-1
3.4×10-1
4.5×10-1
1.7×10-2
2.3×10-2
0.0×10-1
0.0×10-1

(c) V =1.5m/s (d) V =2.0m/s

Fig. 32. The flow field at different velocity
It is seen that the flow field patterns in the figures are nearly the same. There was high flow
rate region near the abrupt curve surfaces of the vehicle head, ballast of the GPS, rudders,
while there was also the low flow field domain on the front of those parts and near the tail
of the vehicle. The high flow rate region area decreases as the velocity increases. The
existence of the mast of GPS makes the flow field behind it disturbed, and makes the flow
field asymmetrical. These changes will increase the drag and hydrodynamic moment on the
vehicle.
The steady turning flow field in longitudinal vertical and horizontal plane with the velocity
of vehicle at 0.5m/s, is shown in figure 33.
It is noted from Figure 33 that the pattern of the steady turning flow field in longitudinal
vertical plane and in horizontal plane has notability difference. Due to the rotational speed,
the flow field is obviously asymmetric and appears large scale high flow rate region and low
flow rate region in the back of the field. An extra hydrodynamic moment is induced because
of the asymmetry of the flow field.
Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 61

Unit：m/s
-1
5.6×10 5.6×10-1
5.0×10-1 5.0×10-1
4.5×10-1 4.5×10-1
3.9×10-1 3.9×10-1
3.7×10-1 3.7×10-1
3.4×10-1 3.4×10-1
2.2×10-1 2.2×10-1
1.7×10-1 1.7×10-1
1.1×10-1 1.1×10-1
5.6×10-2 5.6×10-2
0.0×10-1 0.0×10-1
(a) Steady turning in longitudinal vertical plane (b) Steady turning in horizontal plane
Fig. 33. The steady turning flow field

6.2 Pressure distribution

The pressure distributions on the vehicle at the speeds 0.5m/s and 2m/s when the angle of
attack α is zero are shown in Figure 34 and Figure 35. There has a tendency that the
pressure on the vehicle gradual reduction from head to tail of the vehicle, a high pressure
region on the head and a low pressure region on the tail, which induced the pressure drag
on the vehicle. The pressure of the high pressure region become higher and the low pressure
region become lower with the speed of the vehicle increasing, it means that the pressure
drag on the vehicle increase with the speed increasing. It can be known from the pressure
distribution on the propeller shroud that pressures drag act on the shroud because of there
has higher pressure inside the shroud and lower pressure outside the shroud. The reason for
thus pressure distribution is that the propeller doesn’t rotating in the glide mode which
makes the velocity of flow inside the shroud slower than the outside. So the shroud should
be removed or the profile changed to reduce the drag on the vehicle in glide mode.

Unit：Pascal Unit：Pascal
1.57×10 2 2.45×103
1.34×102 2.06×103
1.11×102 1.66×103
8.77×101 1.27×103
6.45×101 8.75×102
4.12×101 4.80×102
1.80×101 8.61×101
-5.19×100 -3.08×102
-2.84×101 -7.02×102
-5.16×101 -1.10×103
-7.48×101 -1.49×103

Fig. 34. Pressure distribution ( V =0.5m/s) Fig. 35. Pressure distribution ( V =2m/s)
The pressure distributions on the vehicle at the speed 0.5m/s when the angle of attack α
isn’t zero are shown in Figure 36. The pressure distribution on the vehicle isn’t symmetry,
the pressure of front flow surface higher than back flow surface, when glide with an angle of
attack. The wing has the biggest degree of asymmetry of the pressure distribution which
62 Autonomous Underwater Vehicles

makes the wings the main lift generating parts. The asymmetry of the pressure distribution
on the vehicle also induces the hydrodynamic moment on the vehicle.

Unit：Pascal
1.46×102
1.19×102
9.24×101
6.57×101
3.89×101
1.21×101
-1.46×101
-4.14×101
-6.81×101
-9.49×101
-1.22×102

Fig. 36. Pressure distribution ( V =0.5m/s, α = 6 )

7. Conclusions
This chapter focuses on the hydrodynamic effects of the main parts of a hybrid-driven
underwater glider especially in the glide mode, and conducts analysis of the simulation
results of the three main hydrodynamic parts by using the computational fluid dynamics
(CFD) ways. The fluent Inc.’s (Lebanon, New Hampshire) CFD software FLUENT 6.2 was
adopted by this article. The main conclusions are:
It is found that the glide efficiency is most significantly influenced by the chord length while
stability of the vehicle is most remarkably affected by the sweep angle, and the location of
the wings mainly affects glide stability but has little influence on glide efficiency. When the
vehicle glides at about 6°attack angle it has the maximum ratio of lift to drag. The endurance
of the hybrid glider with the same configuration as PETREL will decrease by 10％~35％
compared with the legacy gliders.
For the rudder we design, the angle of stall is about 34° for the three dimensional rudders
and about 20° for the two-dimensional foil section, so the angle of stall of three dimensional
rudder is greater than two-dimension foil section. The area of the rudder of PETREL was
calculated using the DNV rules；The hinge moments are little when P = 0.4c for the rudder
we design no matter how the angle of attack changes.
It was found that overall drag increased by 21 to 26 percent for the model with the propeller
shroud compared with the one without a shroud, but with the same structure and size, the
shroud’s resistance is mainly pressure force. The shroud made the lift-to-drag of the vehicle
in glide mode decrease by as much as 20 percent when the angle of attack was 2º. As the
angle of attack increased, the shroud’s effect was minimized, and the decrease in lift-to-drag
ratio ranged down to five percent at an angle of attack of 20º, meaning glide efficiency
decreased due to the propeller shroud. Finally, the shroud decreases the stability of the
HUG when the angle of attack is lower than the critical angle, but increases it when the
angle of attack is higher than the critical angle. The critical angle is between 8º and 10º for
velocities lower than one meter per second, and between 10ºand 12ºfor velocities in the
range of one to two meters per second.
Hydrodynamic Characteristics of the Main Parts of a Hybrid-Driven Underwater Glider PETREL 63

These findings indicate that the shroud of the underwater glider will increase drag, decrease
the glide efficiency, but it improves the stability when the angle of attack is larger than 8º.
Therefore, the shroud is not a successful design element for the HUG in glide mode, but it
can increase the thrust of the vehicle in propeller mode.
Using CFD to analyze the shroud’s hydrodynamic effects shows that the vehicle should only
be equipped with this feature for activities requiring operation in propeller mode.
Finally, the velocity field, pressure distribution of the hybrid glider PETREL were analyzed,
which make us understand how those main parts effect on the hydrodynamic characteristic
of the vehicle.

8. References
[1] C. C. Eriksen, T. J. Osse, R. D. Light, T, et al, (2001)“Sea glider: A long range autonomous
underwater ve hicle for oceanographic research,” IEEE Journal of Oceanic
Engineering, Vol. 26, 2001, pp. 424–436
[2] J. Sherman, R. E. Davis, W. B. Owens, et al , “The autonomous underwater glider
“Spray”,” IEEE Journal of Oceanic Engineering, vol.26, 2001, pp. 437–446
[3] D. C. Webb, P. J. Simonetti, C. P. Jones, “SLOCUM, an underwater glider propelled by
environmental energy,” IEEE Journal of Oceanic Engineering, vol.26, 2001, pp. 447–
452
[4] R. E. Davis, C. C. Eriksen and C. P. Jones, “Autonomous Buoyancy-driven Underwater
Gliders,” The Technology and Applications of Autonomous Underwater Vehicles.
G.. Griffiths, ed., Taylor and Francis, London, 2002, pp. 37-58
[5] R. Bachmayer, N. E. Leonard, J. Graver, E. Fiorelli, P. Bhatta, D. Paley, “Underwater
gliders: Recent development and future applications,” Proc. IEEE International
Symposium on Underwater Technology (UT'04), Tapei, Taiwan, 2004, pp.195- 200
[6] D. L. Rudnick, R. E. Davis, C. C. Eriksen, D. M. Fratantoni, and M. J. Perry, “Underwater
gliders for ocean research, ” Marine Technology Society Journal, vol. 38, Spring 2004,
pp. 48–59
[7] S. A. Jenkins, D. E. Humphreys, J. Sherman, et al. “Underwater glider system study.”
Technical Report, Office of Naval Research, 2003
[8] Marthiniussen, Roar ; Vestgård, Karstein ; Klepaker, Rolf Arne ; Strkersen, Nils ,Fuel cell
for long-range AUV ,Sea Technology, vol. 38, pp. 69-73
[9] Blidberg, D.Richard ,Solar-powered autonomous undersea vehicles. Sea Technology, vol.
38, pp. 45-51
[10] Ferguson, J , Pope, A.,Butler, B.1, Verrall, R.I.1 ,Theseus AUV-Two record breaking
missions, Sea Technology, vol. 40, pp. 65-70
[11] Albert M. Bradley, Michael D. Feezor, Member, IEEE, Hanumant Singh, and F. Yates
Sorrell, Power Systems for Autonomous Underwater Vehicles, IEEE JOURNAL OF
OCEANIC ENGINEERING, VOL. 26, NO. 4, OCTOBER 2001,pp.526-538
[12] Amit Tyagi*, Debabrata Sen, Calculation of transverse hydrodynamic coefficients using
computational fluid dynamic approach, Ocean Engineering 2006 (33)，pp. 798–809
[13] Douglas E.H, Correlation and validation of a CFD based hydrodynamic &dynamic
model for a towed undersea vehicle, OCEANS, 2001,Vol3, pp.1755-1760
[14] Li Wanping. (2004).Computational fluid dynamics. Huazhong University of Science and
Technology Press, 978-7-5609-3214-9. Wuha (in Chinease)
64 Autonomous Underwater Vehicles

[15] Wu Wangyi. (1983).Fluid mechanics (volume I). Higher Education Press,

7301001991,Beijin. (in Chinease).
[16] Wu Ziniu. (2001) .The basic principles of computational fluid dynamics，science
press，7-03-008128-5, Beijin. (in Chinese)
[17] Graver J G. Underwater gliders: dynamics, control and design. The USA: Princeton
University, 2005
[18] Wu Jianguo, Chen Chaoyin, Wang Shuxin, et al, Hydrodynamic Characteristics of the
Wings of Hybrid-Driven Underwater Glider in Glide Mode, Journal of Tianjin
University, 2010,Vol43,No.1,pp.84-89, (in Chinease)
[19] Jenkins S A, Humphreys D E, Sherman J, et al, Alternatives for enhancement of
transport economy in underwater gliders, IEEE Proceedings of OCEANS, 2003. 948-
950
[20] Wu Baoshan, Xing Fu, Kuang Xiaofeng, et al, Investigation of hydrodynamic
characteristics of submarine moving close to the sea bottom with CFD methods,
Journal of Ship Mechanics, 2005, Vol.9,No.3,pp. 19-28
[21] Shi Shengda. 1995. Submarine Maneuverability. National Defense Industry Press，
9787118013498,Beijing(in Chinese).
[22] Wu Jianguo, Zhang Hongwei, Design and research on the rudder of Mini-type AUV,
ocean technology, 2009，Vol3, No.2, pp.5-8.
[23] P.M.Ostafichuk, AUV hydrodynamics and modeling for improve control, Canada: University
of British Columbia, 2004
[24] Timothy Curtis, B.Eng, The design, conatruction, outfitting, and preliminary testing of
the C-SCOUT autonomous underwater vehicle (AUV), Canada, Faculty of
Engineering and Applied Science Memorial University of Newfoundland, 2001
[25] Jianguo Wu, Chaoying Chen and Shunxin Wang, Hydrodynamic Effects of a shroud
Design For a Hybrid-Driven Underwater Glider,Sea Technology,2010, Vol.51,
No.6,pp.45-47
 Part 2

Navigation and Control

 4

Real-Time Optimal Guidance and

Obstacle Avoidance for UMVs
Oleg A. Yakimenko and Sean P. Kragelund
Naval Postgraduate School Monterey, CA
USA

1. Introduction
The single most important near-term technical challenge of developing an autonomous
capability for unmanned vehicles is to assess and respond appropriately to near-field
objects in the path of travel. For unmanned aerial vehicles (UAVs), that near field may
extend to several nautical miles in all directions, whereas for unmanned ground and
maritime vehicles, the near field may only encompass a few dozen yards directly ahead of
the vehicle. Nevertheless, when developing obstacle avoidance (OA) manoeuvres it is
often necessary to implement a degree of deliberative planning beyond simply altering
the vehicle’s trajectory in a reactive fashion. For unmanned maritime vehicles (UMVs) the
ability to generate near-optimal OA trajectories in real time is especially important when
conducting sidescan sonar surveys in cluttered environments (e.g., a kelp forest or coral
reef), operations in restricted waterways (e.g., rivers or harbours), or performing feature-
based, terrain-relative navigation, to name a few. For example, a primary objective of
sidescan sonar surveys is 100% area coverage while avoiding damage to the survey
vehicle. Ideally, a real-time trajectory generator should minimize deviations from the pre-
planned survey geometry yet also allow the vehicle to retarget areas missed due to
previous OA manoeuvres. Similarly, for operations in restricted waterways, effective OA
trajectories should incorporate all known information about the environment including
terrain, bathymetry, water currents, etc.
In the general case, this OA capability should be incorporated into an onboard planner or
trajectory generator computing optimal (or near-optimal) feasible trajectories faster than in
real time. For unmanned undersea vehicles (UUVs) the planner should be capable of
generating full, three-dimensional (3D) trajectories, however some applications may require
limiting the planner’s output to two-dimensions (2D) for vertical-plane or horizontal-plane
operating modes. For unmanned surface vehicles (USVs) the latter case is the only mode of
operations.
Consider a typical hardware setup consisting of a UUV augmented with an autopilot (Fig.1).
The autopilot not only stabilizes the overall system, but also enables vehicle control at a
higher hierarchical level than simply changing a throttle setting δT (t ) , or deflecting stern
plane δ s (t ) or rudder δ r (t ) angles.
In Fig.1, x WP , y WP , z WP are the vectors defining x, y, and z coordinates of some points in
the local tangent (North-East-Down (NED)) plane for waypoint navigation. Alternatively a
68 Autonomous Underwater Vehicles

typical autopilot may also accept some reference flight-path angle γ (t ) (or altitude/depth)
command and heading Ψ (t ) (or yaw angle ψ (t ) ), respectively. The motion sensors,
accelerometers, and rate gyros measure the components of inertial acceleration, xI (t ) , y I (t )
and zI (t ) , and angular velocity – roll rate p(t ) , pitch rate q(t ) , and yaw rate r (t ) .

xWP , yWP , zWP

Reference Signal δ(t )
γ (t ), Ψ (t ) Controller Vehicle x (t ), y (t ), z (t )
Generator
z (t ),ψ (t )
xI (t ), yI (t ), z I (t )
Autopilot Sensors
p(t ), q(t ), r (t )
Augmented Vehicle

Fig. 1. A. UUV augmented with an autopilot

A trajectory generator would consider an augmented UUV as a new plant (Fig.2) and
provide this plant with the necessary inputs based on the mission objectives (final
destination, time of arrival, measure of performance, etc.). Moreover, the reference signals,
γ (t ) and Ψ(t ) , are to be computed dynamically (once every few seconds) to account for
disturbances (currents, etc.) and newly detected obstacles.

Dynamic Trajectory γ (t ) , Ψ (t )
ref ref
Mission Augmented Vehicle x(t ), y (t ), z (t )
goals Generator z (t ) ,ψ (t )
ref ref
(with Sensors and Controller)

Sensor Data
Position Estimate

Fig. 2. Providing an augmented UUV with a reference trajectory

Ideally, the trajectory generator software should also produce the control inputs δref (t )
corresponding to the feasible reference trajectory (Fig.3) (Basset et al., 2008). This enhanced
setup assures that the inner-loop controller deals only with small errors. (Of course this
setup is only viable if the autopilot accepts these direct actuator inputs.)

δ ref (t )
Dynamic Trajectory γ (t ) , Ψ (t )
ref ref
Mission Augmented Vehicle x(t ), y (t ), z (t )
goals Generator z (t ) ,ψ (t )
ref ref (with Sensors and Controller)

Sensor Data
Position Estimate

Fig. 3. Providing an augmented UUV with the reference trajectory and reference controls
The goal of this chapter is to present the dynamic trajectory generator developed at the
Naval Postgraduate School (NPS) for the UMVs of the Center for Autonomous Vehicle
Research (CAVR) and show how the OA framework is built upon it. Specifically, Section 2
formulates a general feasible trajectory generation problem, followed by Section 3, which
introduces the general ideas behind the proposed framework for solving this problem that
utilizes the inverse dynamics in the virtual domain (IDVD) method. Section 4 considers
simplifications that follow from reducing the general spatial problem to two planar
subcases. Section 5 describes the REMUS UUV and SeaFox USV and their forward looking
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 69

sonar (FLS) systems employed for OA research at NPS. Section 6 addresses path-following
considerations and practical implementation details for tracking nonlinear trajectories with
conventional vehicle autopilots. Section 7 presents results from computer simulations and
field experiments for several different scenarios which benefit from faster-than-real-time
computation of near-optimal trajectories.

2. Problem formulation
Let us consider the most general case and formulate an optimization problem for computing
collision-free trajectories in 3D (it can always be reduced to a 2D problem by eliminating
two states). We will be searching within a set of admissible trajectories described by the state
vector

{ }
z(t ) = [ x(t ), y(t ), z(t ), u(t ), v(t ), w(t )] ∈ S , S = z(t ) ∈ Z6 ⊂ E6 , t ∈ ⎡⎣t0 , t f ⎤⎦
T
(1)

where the components of the velocity vector – surge u, sway v, and heave w, defined in the
body frame {b} – are added to the UUV NED coordinates x, y and z ( z = 0 at the surface and
increases in magnitude with depth). While many UUVs are typically programmed to
operate at a constant altitude above the ocean floor, it is still preferable to generate vertical
trajectories in the NED local tangent plane because the water surface is a more reliable
absolute reference datum than a possibly uneven sea floor. In general, however, it is a trivial
matter to convert the resulting depth trajectory z(t) to an altitude trajectory h(t) for vehicles
equipped with both altitude and depth sensors. Section 6.2 describes such practical
considerations in detail.
The admissible trajectories should satisfy the set of ordinary differential equations
describing the UUV kinematics

⎡ x( t ) ⎤ ⎡ u(t ) ⎤
⎢ ⎥ u ⎢ ⎥
⎢ y ( t ) =
⎥ b ⎢ v(t ) ⎥
R (2)
⎢⎣ z(t ) ⎥⎦ ⎢⎣ w(t )⎥⎦

In (2) ub R is the rotation matrix from the body frame {b} to the NED frame {u}, defined using
two Euler angles, pitch θ (t ) and yaw ψ(t ) , and neglecting a roll angle as

⎡ cos ψ(t )cos θ(t ) − sin ψ(t ) cos ψ(t )sin θ(t )⎤
u ⎢ ⎥
b R(t ) = ⎢ sin ψ(t )cos θ(t ) cos ψ(t ) sin ψ(t )sin θ(t ) ⎥ (3)
⎢⎣ − sin θ(t ) 0 cos θ(t ) ⎥⎦

Although we are not going to exploit it in this study, admissible trajectories should also
obey UUV dynamic equations describing translational and rotational motion. This means
that the following linearized system holds for the vector ς(t), which includes speed
components u, v, w (being a part of our state vector z(t)) and angular rates p, q, r:

ς(t ) = Aς(t ) + Bδ(t ) (4)

Here A and B are the state and control matrices and δ = [δT , δ s , δ r ]T is the control vector
(Healey, 2004).
70 Autonomous Underwater Vehicles

Next, the admissible trajectories (1) should satisfy the initial and terminal conditions

z (t0 ) = z 0 , z(t f ) = z f (5)

Finally, certain constraints should be obeyed by the state variables, controls and their
derivatives. For example, in the case of a UUV these can include obvious constraints on the
UUV depth:

zmin ≤ z(t ) ≤ zmax (6)

where zmax ( x , y ) describes a programmed operational depth limit. For vehicles

programmed to operate at some nominal altitude above the sea floor, the zmax ( x , y )
constraint can be converted into a minimum altitude hmin ( x , y ) constraint as described in
Section 6.2.
A 3D OA requirement can be formulated as

[ x(t ); y(t ); z(t )] ∩ ℜ = 0 (7)

where ℜ is the set of all known obstacle locations. The constraints are usually imposed not
only on the controls themselves δ ≤ δmax but on their time derivatives as well δ ≤ δmax to
account for actuator dynamics. Knowing the system’s dynamics (4) (or simply complying
with the autopilot specifications), these latter constraints can be elevated to the level of the
reference signals, for instance

θ (t ) ≤ θmax and ψ (t ) ≤ ψ max (8)

The objective is to find the best trajectory and corresponding control inputs that minimize
some performance index J. Typical performance index specifications include: i) minimizing
time of the manoeuvre t f − t0 , ii) minimizing the distance travelled to avoid the obstacle(s),
and iii) minimizing control effort or energy consumption. In addition, the performance
index may include some “exotic” constraints dictated by a sensor payload. For example, a
UUV may require vehicle trajectories which point a fixed FLS at specified terrain features or
minimize vehicle pitch motion in order to maintain level, horizontal flight along a survey
track line for accurate synthetic aperture sonar imagery (Horner et al., 2009).
Before we proceed with the development of the control algorithm, it should be noted that
quite often the UUV surge velocity is assumed to be constant, u(t ) ≡ U 0 , to provide enough
control authority in two other channels. This uniquely defines a throttle setting δT (t ) , and
leaves only two control inputs, δ s (t ) and δ r (t ) , for altering the vehicle’s trajectory. It also
allows us to consider matrices A and B in (4) to be constant (time- and states-independent).
If this assumption is not required, inverting kinematic and dynamic equations will differ
slightly from the examples presented in the next section. However, the general ideas of the
proposed approach remain unchanged.

3. Real-time near-optimal guidance

For the dynamic trajectory generator shown in Figs. 2 and 3, it is advocated to use the direct-
method-based IDVD (Yakimenko, 2000). The primary rationale is that this approach features
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 71

several important properties required for real-time implementations: i) the boundary

conditions including high-order derivatives are satisfied a priori; ii) the resulting control
commands are smooth and physically realizable, iii) the method is very robust and is not
sensitive to small variations in input parameters, iv) any compound performance index can
be used during optimization. Moreover, this specific method uses only a few variable
parameters, thus ensuring that the iterative process during optimization converges very fast
compared to other direct methods. The IDVD-based trajectory generator consists of several
blocks. The goal of the first block, to be discussed next, is to produce a candidate trajectory,
which satisfies the boundary conditions.

3.1 Generating a candidate trajectory

Again, consider the most general case of a UUV operating in 3D (as opposed to a USV).
Suppose that each coordinate xi , i = 1, 2, 3 of the candidate UUV trajectory is represented as
a polynomial of degree M of some abstract argument τ , the virtual arc

M
xi (τ ) = ∑ aikτ k , (9)
k =0

(for simplicity of notation we assume x1 (τ ) ≡ x(τ ) , x2 (τ ) ≡ y (τ ) and x3 (τ ) ≡ z (τ ) ). In general,

analytic expressions for the trajectory coordinates can be constructed from any combination
of basis functions to produce a rich variety of candidate trajectories. For example, a
combination of monomials and trigonometric functions was utilized in (Yakimenko &
Slegers, 2009).
As discussed in (Yakimenko, 2000; Horner & Yakimenko, 2007) the degree M is determined
by the number of boundary conditions that must be satisfied. Specifically, it should be
greater or equal to the number of preset boundary conditions but one. In general the desired
trajectory includes constraints on the initial and final position, velocity and acceleration: xi 0 ,
xif , xi′0 , xif′ , xi′′0 , xif′′ . In this case the minimal order of polynomials (9) is 5, because all
coefficients in (9) will be uniquely defined by these boundary conditions, leaving the
“length” of the virtual arc τ f as the only varied parameter. For more flexibility in the
candidate trajectory, additional varied parameters can be obtained by increasing the order
of the polynomials (9). For instance, using seventh-order polynomials will introduce two
more varied parameters for each coordinate expression. Rather than varying two coefficients
in these extended polynomials directly, we will vary the initial and final jerk, xi′′′0 and xif′′′ ,
respectively. In this case, coefficients aik in (9) can be determined by solving the obvious
system of linear algebraic equations equating polynomials (9) to xi 0 , xif , xi′0 , xif′ , xi′′0 , xif′′ ,
xi′′′0 and xif′′′ at two endpoints ( τ = 0 and τ = τ f ) (Yakimenko, 2000, 2008).
By construction, the boundary conditions (5) will be satisfied unconditionally for any value
of the final arc τ f . However, varying τ f will alter the shape of the candidate trajectory.
Figure 4 demonstrates a simple example whereby a UUV operating 2m above the sea floor
at 1.5m/s must perform a pop-up manoeuvre to avoid some obstacle. Even with a single
varied parameter, changing the value of τ f allows the UUV to avoid obstacles of different
heights. Similar trajectories could be produced solely in the horizontal plane or in all three
dimensions. It should be pointed out that even at this stage infeasible candidate trajectories
72 Autonomous Underwater Vehicles

will be ruled out. (In Fig.4 the trajectory requiring the UUV to jump out of the water is
infeasible because it violates the constraint (6).)

Fig. 4. Varying the candidate trajectory while changing the value of τ f

With six free parameters, which in our case are components of the initial and final jerk ( xi′′′0
and xif′′′ , i = 1, 2, 3 ) the trajectory generator can change the overall shape of the trajectory
even further. To this end, Fig.5 illustrates candidate trajectories for a UUV avoiding a 10m
obstacle located between its initial and final points. These trajectories were generated by
varying just two components of the jerk, x30 ′′′ and x3′′′f , and minimizing τ f . This additional
flexibility can produce trajectories which satisfy operational constraints (6), as well as OA
constraints (7).

Fig. 5. Candidate trajectories obtained by varying the terminal jerks

The selection of a specific trajectory will be based upon whether the trajectory is feasible
(satisfies constraints (8)) and if so, whether it assures the minimum value of the performance
index calculated using the values of the vehicle states (and controls) along that trajectory. As
an example, Fig.6 presents collision-free solutions for two different locations of a 10m-tall
obstacle when five varied parameters, x10 ′′′ , x1′′′f , x30
′′′ , x3′′′f and τ f , are optimized to assure
feasible minimum-path-length trajectories

Fig. 6. Examples of minimum-path-length trajectories

Now, let us address the reason for choosing some abstract parameter τ as an argument for
the reference functions (9) rather than time or path length, which are commonly used.
Assume for a moment that τ ≡ t . In this case, once we determine the trajectory we
unambiguously define a speed profile along this trajectory as well, since
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 73

V (t ) = u(t )2 + v(t )2 + w(t )2 = x(t )2 + y(t )2 + z(t )2 (10)

Obviously, we cannot allow this to happen because we want to vary the speed profile
independently. With the abstract argument τ this becomes possible via introduction of a
speed factor λ such that

dτ
λ (τ ) = . (11)
dt
Now instead of (10) we have

V (τ ) = λ (τ ) x′(τ )2 + y′(τ )2 + z′(τ )2 (12)

and by varying λ (τ ) we can achieve any desired speed profile.

The capability to satisfy higher-order derivatives at the trajectory endpoints, specifically at
the initial point, allows continuous regeneration of the trajectory to accommodate sudden
changes like newly discovered obstacles. As an example, Fig.7 demonstrates a scenario
whereby a UUV executing an OA manoeuvre discovers a second obstacle and must generate
a new trajectory beginning with the current vehicle states and control values (up to the
second-order derivatives of the states). The suggested approach enables this type of
continuous trajectory generation and ensures smooth, non-shock transitions.

Fig. 7. Example of dynamic trajectory reconfiguration

3.2 Inverse dynamics

The second key block inside the dynamic trajectory generator in Figs. 2 and 3 accepts the
candidate trajectory as an input and computes the components of the state vector and
control signals required to follow it. In this way we can ensure that each candidate trajectory
does not violate any constraints (including those of (8)).
First, using the following relation for any parameter ζ ,

dζ dτ
ζ (τ ) = = ζ ′(τ )λ (τ ) (13)
dτ dt
we convert kinematic equations (2) into the τ domain

⎡ x′(τ )⎤ ⎡ U0 ⎤
λ (τ ) ⎢ y (τ )⎥ = b R ⎢ v(τ ) ⎥⎥
⎢ ′ ⎥ u ⎢ (14)
⎢⎣ z′(τ ) ⎥⎦ ⎢⎣ w(τ )⎥⎦

Next, we assume the pitch angle to be small enough to let sin θ (t ) ≈ 0 and cosθ (t ) ≈ 1 , so
that the rotation matrix (3) becomes
74 Autonomous Underwater Vehicles

⎡cos ψ (τ ) − sin ψ (τ ) 0 ⎤
u
b R(τ ) = ⎢⎢ sin ψ (τ ) cos ψ (τ ) 0 ⎥⎥ (15)
⎢⎣ 0 0 1 ⎥⎦

While this step is not required, it simplifies the expressions in the following development.
Inverting (14) via the rotation matrix (15) yields

⎡U 0 ⎤ ⎡ cos ψ sin ψ 0 ⎤ ⎡ x′ ⎤
⎢ ⎥ ⎢ ⎥ ⎢ ′⎥
⎢ v ⎥ = λ ⎢ − sin ψ cos ψ 0 ⎥ ⎢ y ⎥ (16)
⎢⎣ w ⎥⎦ ⎢⎣ 0 0 1 ⎥⎦ ⎢⎣ z′ ⎥⎦

Hereafter each variable’s explicit dependence on τ will be omitted from the notation.
Now the three equations of system (16) must be resolved with respect to three unknown
parameters, v, w and ψ. While the last one readily yields

w = λ z′ (17)
the first two require more rigorous analysis.
Consider Fig.8. Geometrically, a scalar product of two vectors on the right-hand-side of the
first equation in (16) represents the length of the longest side of the shaded rectangle.
Similarly, the second equation expresses the length of the shortest side of this rectangle.
From here it follows that the square of the length of the diagonal vector can be expressed in
two ways: v 2 λ −2 + U 02 λ −2 = x′2 + y′2 . This yields

v = λ 2 ( x′2 + y′2 ) − U 02 (18)

From the same figure it follows that

vλ −1 v y′
ψ = Ψ − tan −1 = Ψ − tan −1 , Ψ = tan −1 (19)
U 0λ −1
U0 x′

U 0 λ −1 ⎡ x′ ⎤
⎢ y′⎥
⎣ ⎦
⎡ cos ψ ⎤
vλ −1 ⎢ sin ψ ⎥
⎣ ⎦
ψ Ψ
y′
Ψ =tan −1
x′
Fig. 8. Kinematics of horizontal plane parameters
Now, using these inverted kinematic equations, we can check whether each candidate
trajectory obeys the constraints imposed on it (constraints (8)).
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 75

3.3 Discretization
We proceed with computing the remaining states along the reference trajectory over a fixed
set of N points (for instance, N=100) spaced evenly along the virtual arc [0;τ f ] with the
interval

Δτ = τ f ( N − 1)−1 (20)
so that

τ j = τ j − 1 + Δτ , j = 2,..., N , ( τ 1 = 0 ) (21)

In order to determine coefficients for polynomials (9) we will have to guess on the values of
the varied parameters τ f , xi′′′0 , x′′′i 0 , xif′′′ , and x′′′if . These guesses will be used along with the
known or desired boundary conditions xi 0 , x′i 0 , xi′′0 , xif , xif′ , and x′′if . The boundary
conditions on coordinates xi 0 and xif come directly from (5). According to (14), the given
boundary conditions on surge, sway, and heave velocities define the first-order time
derivatives of the coordinates as

⎡ x′0; f ⎤ ⎡U ⎤
⎢ ⎥ −1 u
⎢ 0 ⎥
⎢ y′0; f ⎥ = λ0; f b R0; f ⎢ v0; f ⎥ (22)
⎢ ⎥ ⎢ ⎥
′
⎣⎢ z0; f ⎦⎥ ⎣⎢ w0; f ⎦⎥
u
They also define the initial and final pitch and yaw angles used to compute b R0; f in (22) as

− w0; f v0; f
θ 0; f = γ 0 + tan −1 and ψ 0; f = Ψ 0 − tan −1 (23)
U02 + v0;
2 U0
f

In equation (22) we may use any value for the initial and final speed factor λ , for example,
λ0; f = 1 . This value simply scales the virtual domain; the higher the values for λ , the larger
the values for τ f . This follows directly from equations (11) and (12) that λ0; f τ −f 1 = U0 s −f 1 ,
where s f is the physical path length.
Finally, initial values for the second-order derivatives are provided by the UUV motion
sensors (see Figs. 1-3) (after conversion to the τ domain), while final values for the second-
order derivatives are usually set to zero for a smooth arrival at the final point. Having an
analytical representation of the candidate trajectory (9) defines the values of xij , and x′ij ,
i = 1, 2, 3 , j = 1,..., N .
Now, for each node j = 1,..., N we compute

λ j = ΔτΔt −j −11 (24)

where

( x j − x j − 1 )2 + ( y j − y j − 1 )2 + ( z j − z j − 1 )2
Δt j − 1 = (25)
U 02 + v 2j − 1 + w 2j − 1
76 Autonomous Underwater Vehicles

and then use (17)-(19) to compute w, v, ψ and Ψ at each timestamp. The vertical plane
parameters, flight path angle γ and pitch angle θ, can be computed using the following
relations:

− z′j −w j
γ j = tan −1 , θ j ≈ γ j + tan −1 (26)
x′j2 + y′j2 U02 + v 2j

In order to check the yaw rate constraints (8) we must first numerically differentiate the
expression for Ψ in (19).

3.4 Optimization
When all parameters (states and controls) are computed in each of N points, we can
compute the performance index J and the penalty function. For example, we can combine
constraints (6) and (8) into the joint penalty

⎡ min(0; z j − zmin )2 ⎤
⎢ j ⎥
⎢ 2 ⎥
⎢ max(0; z j − zmax ) ⎥
j
Δ = ⎣⎡ k zmin , k zmax , kθ , kψ ⎦⎤ ⎢ ⎥ (27)
⎢ max(0; θ j − θ max )2 ⎥
⎢ j ⎥
⎢ 2⎥
max(0;
⎢ j ψ j − ψ )
max ⎥
⎣ ⎦
with k zmin , k zmax , k θ and kψ being scaling (weighting) coefficients. Now the problem can be
solved using numerical methods such as the built-in fmincon function in the Mathworks’
MATLAB development environment. Alternatively, by combining the performance index J
with the joint penalty Δ we may exploit MATLAB’s non-gradient fminsearch function. For
real-time applications, however, the authors prefer to use a more robust optimization routine
based on the gradient-free Hooke-Jeeves pattern search algorithm (Yakimenko, 2011).

4. Planar cases
This section presents two simplified cases for a vehicle manoeuvring exclusively in either
the horizontal or vertical plane.

4.1 Horizontal plane guidance

For the case of a UUV manoeuvring in the horizontal plane or a USV, the computational
procedure is simplified. The trajectory is represented by only two reference polynomials for
coordinates x1 and x2. Hence, we end up having only five varied parameters, which
′′′ , x′′′20 , x1′′′f and x′′′
are: τ f , x10 2 f . The remaining kinematic formulas are identical to those
presented above with z ≡ 0 , z′ ≡ 0 and γ ≡ 0 . Figure 9 shows an example of a planar
scenario in which a USV has to compute a new trajectory twice. First, after detecting an
obstacle blocking its original path, a new trajectory is generated to steer right and pass
safely in front of the object (dotted line). Second, while executing the first avoidance
manoeuvre the USV detects that the object has moved south into its path. It therefore
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 77

generates a new trajectory to steer left and pass safely behind the object’s stern. The
complete trajectory is shown as a solid line.

4.2 Vertical plane guidance

For the case of a UUV manoeuvring in the vertical plane, the 3D algorithm can be reduced
to the 2D case in a manner similar to the horizontal case. Specifically, using five varied
parameters, τ f , x10 ′′′ , x1′′′f and x3′′′f , we can develop reference trajectories for x1 and x3,
′′′ , x30
and then use the same general equations developed in Section 3, assuming y ≡ 0 , y′ ≡ 0 ,
and Ψ ≡ 0 .

Fig. 9. Moving obstacle avoidance in a horizontal plane

Alternatively, we can use a single reference polynomial to approximate just x3 and then
integrate the third equation of (4) to get the heave velocity w . That allows computation of
the time period Δt j − 1 using

Δt j − 1 = ( z j − z j − 1 )w−j −11 (28)

instead of (25).
Another way of dealing with vertical plane manoeuvres is to invert the dynamic equations
(4) (Horner & Yakimenko, 2007). After developing the reference functions for two
coordinates, x1 and x3 , the stern plane δ s control input is computed subject to five variable
parameters: τ f , x10 ′′′ , x1′′′f , and x3′′′f .
′′′ , x30
In this case, the corresponding time period Δt j − 1 is computed similarly to (28):

z j − z j −1 z j − z j −1
Δt j − 1 = t j − t j − 1 = ≈ (29)
w j − 1 cosθ j − 1 + u0 sin θ j − 1 w j −1

and the heave velocity is calculated using the third equation of system (4) as

x′j − 1 = λ j−−11 ( w j − 1 sin θ j − 1 + U0 cosθ j − 1 ) x j = x j − 1 + Δτ x′j − 1

(30)
(
w′j − 1 = λ j−−11 A 33 w j − 1 + A 35q j − 1 + B 32δ s ; j − 1 ) w j = w j − 1 + Δτ w′j − 1

The next step involves computing the pitch angle, pitch rate and pitch acceleration as

⎛ u0 x′j + w j z′j ⎞ θ − θ j −1 q j − q j −1
θ j = cos−1 ⎜ λ j ⎟, qj = θj ≈ j , qj = θj ≈ (31)
⎜ 2
w j + U0 ⎠2 ⎟
Δt j − 1 Δt j − 1
⎝
Finally, we can compute the dive plane deflection required to follow the trajectory using the
5th equation of system (4) as
78 Autonomous Underwater Vehicles

−1
δ s ; j = (q j − A53 w j − A55q j )B52 (32)

In this case the last two terms in the joint penalty Δ , similar to that of (27) but developed for
the new controls, enforce δ s ≤ δ s max and δ s ≤ δ s max .

5. Test vehicles and sensing architecture

The preceding trajectory generation framework has been implemented on several UMVs.
Before presenting simulated and experimental results with specific vehicle platforms at sea,
we first introduce two such vehicles in use at CAVR - the REMUS UUV and SeaFox USV.
Both vehicles utilize FLS to detect and localize obstacles in their environment and employ
the suggested direct method to generate real-time OA trajectories. This section provides
system-level descriptions of both platforms including their sensors, and proposes a way of
building the OA framework on top of the trajectory generation framework, i.e. enhancing
the architecture of Figs. 2 and 3 even further.

5.1 REMUS UUV and SeaFox USV

Remote Environmental Monitoring UnitS (REMUS) are UUVs developed by Woods Hole
Oceanographic Institute and sold commercially by Hydroid, LLC (Hydroid, 2011). The NPS
CAVR owns and operates two REMUS 100 vehicles in support of various navy-sponsored
research programs. The REMUS 100 is a modular, 0.2m diameter UUV designed for
operations in coastal environments up to 100m deep. Typical configurations measure less
than 1.6m in length and weigh less than 45kg, allowing the entire system to be easily
transported worldwide and deployed by a two-man team (Fig.10a). Designed primarily for
hydrographic surveys, REMUS comes equipped with sidescan sonar and sensors for
collecting oceanographic data such as conductivity, temperature, depth or optical
backscatter. The REMUS 100 system navigates using a pair of external transponders for long
baseline acoustic localization or ultra-short baseline terminal homing, as well as an Acoustic
Doppler Current Profiler/Doppler Velocity Log (ADCP/DVL). The ADCP/DVL measures
vehicle altitude, ground- or water-relative vehicle velocity, and current velocity profiles in
body-fixed coordinates.

a) b)
Fig. 10. NPS REMUS 100 UUV (a) and FLS arrays (b)
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 79

To support ongoing CAVR research into sonar-based OA, terrain-relative navigation, and
multi-vehicle operations in cluttered environments, each NPS REMUS vehicle has been
modified to incorporate a FLS, multi-beam bathymetric sonar, acoustic communications
modem, navigation-grade inertial measurement system, and fore/aft horizontal/vertical
cross-body thrusters for hovering or precise manoeuvring. (Figure 10b provides a close up
of the NPS REMUS FLS arrays with nose cap removed.) To maximize the REMUS system’s
utility as a research platform, Hydroid developed the RECON communications interface so
that sensor and computer payloads can interact with the REMUS autopilot. Using this
interface, NPS payloads receive vehicle sensor data and generate autopilot commands based
on NPS sonar processing, trajectory generation, and path-following algorithms.
The SeaFox USV was designed and manufactured by Northwind Marine (Seattle, WA) as a
remote-controlled platform for intelligence, surveillance, reconnaissance, anti-terrorist force
protection, and maritime interdiction operations (Northwind Marine, 2011). SeaFox is a
4.88m long, aluminium, rigid-hulled inflatable boat with a 1.75m beam; 0.25m draft; fold-
down communications mast; and fully-enclosed electronics and engine compartments.
SeaFox’s water jet propulsion system is powered by a JP5-fueled, 185-HP V-6 Mercury
Racing engine, and can deliver a top speed of 74km/h. Standard sensing systems include
three daylight and three low light navigation cameras for remote operation, as well as twin
daylight and infrared gyro-stabilized camera turrets for video surveillance. All video is
accessible over a wireless network via two onboard video servers.
The NPS SeaFox was modified to enable fully-autonomous operations by integrating a
payload computer with the primary autopilot (Fig.11). Meanwhile, the original remote
control link was retained to provide an emergency stop function. NPS algorithms running
on the payload computer generate rudder and throttle commands that are sent directly to
the SeaFox autopilot. Recent navigational upgrades include a satellite compass that uses
differential Global Positioning System (GPS) navigation service for accurate heading
information, a tactical-grade inertial measurement unit for precise attitude estimation, and
an optional ADCP/DVL for water velocity measurements. To support ongoing CAVR
research into autonomous riverine navigation, the NPS SeaFox was further upgraded to
deploy a retractable, pole-mounted FLS system for underwater obstacle detection and
avoidance (Gadre et al., 2009). Figure 12 shows the SeaFox USV operating on a river with its
sonar system deployed below the waterline.

5.2 Sonar system

The NPS REMUS and SeaFox vehicles rely on FLS to detect and localize obstacles in their
environment. Both platforms utilize commercial blazed array sonar systems manufactured
by BlueView Technologies (BlueView Technologies, 2011). These sonar systems comprise
one or more pairs of arrays grouped into sonar “heads.” Each sonar head generates a 2D
cross-sectional image of the water column in polar coordinates, typically plotted as the
image plane’s field of view angle vs. range. Due to the sonar arrays’ beam width, the
resulting FLS imagery has a 12-degree out-of-plane ambiguity. The REMUS FLS system
consists of two fixed sonar heads, which provide a 90-degree horizontal field of view (FOV)
and a 45-degree vertical FOV. Similarly, the SeaFox FLS system is comprised of twin sonar
heads mounted on port and starboard pan/tilt actuators, providing each side with a 45-
degree FOV image at an adjustable mounting orientation that can be swept through the
water column for increased sensor coverage.
80 Autonomous Underwater Vehicles

Water Speed DVL (Optional)

Sensor 440MHz

Remote
E-Stop
DGPS Compass RS-232
RS-232
Radio
Process 1
Modem
RS-232 Process 2

RS-422 MOOS RS-232 SeaFox

Process N
DB Switch Autopilot
RS-485
Throttle and Rudder
HG1700 Commands
Secondary Controller
Pan and Tilt
Commands

IMU RS-485
PC/104

2.4GHz

Starboard Port
Sonar Head Sonar Head
Ethernet Wave Relay
Switch MANET

Fig. 11. SeaFox sensors and control architecture

Fig. 12. SeaFox USV navigating on the Sacramento River near Rio Vista, CA

5.3 Obstacle avoidance framework

The proposed OA framework built into the architecture of Figs. 2 and 3 is shown in Fig.13. It
consists of an environmental map, a planning module, a localization module, sensors and
actuators (Horner & Yakimenko, 2007). The environmental map can include a priori
knowledge, such as the positions of charted underwater obstacles, but also incorporate
unexpected threats discovered by sonar. The positions of all obstacles are eventually
resolved in the vehicle-centred coordinate frame with the help of the localization module.
The planning module is responsible for generating collision-free trajectories the vehicle
should follow. This reference trajectory, possibly with reference controls, is then used to
excite actuators.
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 81

Mission goals

Dynamic Trajectory Generator

Environmental
Deliberative
Map

Localization Reactive

Obstacle Avoidance Framework

Sensor Data Augmented Vehicle Reference Trajectory

Position Estimate (with Sensors and Controller) Reference Controls

Fig. 13. Components of the NPS OA framework

The proposed OA framework supports both deliberative and reactive obstacle avoidance
behaviours. Deliberative OA involves the ability to generate and follow a trajectory that
avoids all known obstacles between an arbitrary start location and some desired goal
location, whereas reactive OA involves the ability to avoid any previously unknown
obstacles detected while following this trajectory. Since the sonar system continuously
resamples the environment, this reactive behaviour can be achieved by a deliberative
planner as long as i) it executes fast enough to incorporate all new obstacle information from
the sonar, and ii) it generates feasible trajectories which begin with the vehicle’s current state
vector. Specifically, since the REMUS and SeaFox FLS have limited range and limited fields
of view in both image planes, new trajectories must be generated continuously (e.g. on some
fixed time interval or upon detection of a new obstacle) during execution of the current
manoeuvre to ensure reactive avoidance of new obstacles.
As an example of deliberative OA, assume a REMUS vehicle is mapping a minefield with
sidescan sonar prior to a mine clearance operation. For this mission, the goal locations are
provided by the sequence of waypoints making up a typical lawn-mowing survey pattern. If
an obstacle is detected along a specified track line, the preferred OA manoeuvre for this
mission would be one that also minimizes the cumulative deviation from this track line,
since we desire 100% sensor coverage of the survey area. Hence, deliberative OA implies the
optimization of some performance index. Likewise, while digital nautical charts or previous
vehicle surveys can be used to identify some obstacles a priori, this data is usually
incomplete or outdated. Vehicles should be capable of storing in memory the locations of
any uncharted obstacles discovered during their mission so that subsequent trajectories can
avoid them—even when they are no longer in the sonar’s current field of view. Deliberative
OA, therefore, also entails the creation and maintenance of obstacle maps.

5.4 Obstacle detection and mapping

Detecting obstacles from sonar imagery is challenging because several factors affect the
intensity of sonar reflections off objects in the water column. These factors include the size,
material, and geometry of an object relative to the sonar head; interference from other
acoustic sensors; and the composition of the acoustic background (e.g. bottom type, amount
of sediment, etc.) to name a few (Masek, 2008). Once an obstacle has been detected, other
image processing algorithms must measure its size and compute its location within the
navigational reference frame. While localizing obstacles via the range and bearing data
82 Autonomous Underwater Vehicles

embedded in the sonar imagery is straightforward, computing their true size is very
difficult. First, for the REMUS FLS, an obstacle’s height and width can be measured directly
by both sonar heads only when it is located within a narrow 12-degree by 12-degree
“window” directly ahead of the vehicle. Due to this narrow beam width, most obstacles are
not imaged by both the horizontal and vertical sonar at the same time. Moreover, FLS
images do not contain information in the region behind an obstacle’s ensonified leading
edge; this portion of the image is occluded. Therefore, the true horizontal and vertical extent
of each obstacle must be deduced from multiple views of the same object. For a vehicle with
a fixed sensor like the REMUS, this may be accomplished by deliberately inducing vehicle
motion to vary the sonar angle (Furukawa, 2006) or by generating trajectories that will
image the object from a different location at a later time. For these scenarios, it is desirable to
balance OA behaviours with exploration behaviours in order to maximize sensor coverage
and generate more complete obstacle maps. In this way, the proposed trajectory generation
framework can be adapted to produce exploratory trajectories which more accurately
measure the size and extent of detected obstacles (Horner et al., 2009). Nevertheless, due to
the uncertainty in sonar images arising from environmental factors, sensor geometry, or
obstacle occlusion, it is prudent to make conservative assumptions about an obstacle’s
boundaries until other information becomes available.
For the remainder of this section, we highlight different representations for incorporating
obstacle size, location, and uncertainty into an obstacle map for efficient collision detection
during the trajectory optimization phase. These representations can be tailored to the
working environment. For operations in a kelp forest, for example, kelp stalks often appear
as point-like features in horizontal-plane sonar imagery (Fig.14) but seldom appear in
vertical-plane images. By making the reasonable assumption (for this environment) that
these obstacles extend vertically from the sea floor to the surface, it may be simpler to
perform horizontal-plane OA through this type of obstacle field. Nevertheless, when
building an obstacle map comprised primarily of point features, mapping algorithms must
account for the uncertainty inherent in sonar imagery. One simple but effective technique
adds spherical (3D) or circular (2D) uncertainty bounds to each point feature stored in the
obstacle map. Candidate OA trajectories which penetrate these boundaries violate constraint
(7). Under this construct, collision detection calculations are reduced to a simple test to
determine whether line segments in a discretized trajectory intersect with the uncertainty
circle (2D) or sphere (3D) for each obstacle in the map. In general, when checking for line
segment intersections with a circle or sphere there are five different test cases to consider
(Bourke, 1992). Our application, however, requires only two computationally efficient tests
to determine: i) which line segment along a discretized trajectory contains the closest point
of approach (CPA) to an obstacle, and ii) whether this CPA is located inside the obstacle’s
uncertainty bound.
Most objects appear in sonar imagery not as point features, but as complex shapes. Unlike
point features, it is difficult and computationally expensive to determine exhaustively
whether or not a candidate vehicle trajectory will collide with these shapes. Instead, we can
bound an arbitrary shape with a minimal area rectangle (or box, in 3D) aligned with the
shape’s principle axes (Fig.15). This type of object, called an oriented bounding box, is
widely used in collision-detection algorithms for video games. One technique, based on the
Separating Axis Theorem from complex geometry, results in an extremely fast test for line
segment intersections with an oriented bounding box (Kreuzer, 2006). With slight
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 83

modification, this test can also be used to detect when a trajectory passes directly above a
bounding box. In our application, we use the OpenCV computer vision library to generate a
bounding box around each object detected in the horizontal image plane. For each box, we
then compute its centre point, length extent, and angle relative to the vehicle’s navigation
frame. Due to occlusion, the width extent produced from this rectangle does not accurately
convey the true size of the obstacle, so we assume a constant value for this parameter. To
create a 3D (actually 2.5D) bounding box around the object, we compute its vertical extents
from vertical sonar imagery. At this time, the assumption is that obstacles extend from the
ocean floor to its measured height above bottom, but this method can be generalized to
obstacles suspended in the water column or extending from the surface to a measured depth
(i.e. ships in a harbour).

Fig. 14. Horizontal FLS image of a kelp forest

Fig. 15. Example of the bounding boxes used in conservative collision detection calculations
While oriented bounding boxes work well for mapping discrete obstacles in open-water
environments, they require an additional image processing step and are not easily adapted
to operations in restricted waterways. For these environments, a probabilistic occupancy
grid is preferable for robustly mapping large continuous obstacles (e.g. harbour
breakwaters) or natural terrain (e.g., a river’s banks). Occupancy grids divide the
environment into a grid of cells and assign each cell a probability of being occupied by an
obstacle. Given a probabilistic sensor model, Bayes’ Theorem is used to compute the
probability that a given cell is occupied, based upon current sensor data. By extension, an
84 Autonomous Underwater Vehicles

estimate for the occupancy state of each cell can be continually updated using an iterative
technique that incorporates all previous measurements (Elfes, 1989). Figure 16a shows an
occupancy grid map of a river generated by the SeaFox FLS system. In this image, each pixel
corresponds to a 1-metre square grid cell whose colour represents the cell’s probability of
being occupied (red) or empty (green). For comparison, the inset portion of the occupancy
grid map has been overlaid with an obstacle map of oriented bounding boxes in Fig.16b.
Clearly, using discrete bounding boxes to represent a long, continuous shoreline quickly
becomes intractable as more and more boxes are required. The occupancy grid framework is
a much more efficient obstacle map representation for wide area operations in restricted
waterways.

a) b)
Fig. 16. Occupancy grid for a river as generated by the SeaFox FLS system
NPS has developed probabilistic sonar models for the BlueView FLS and has successfully
combined separate 2D occupancy grids in order to reconstruct the 3D geometry of an
obstacle imaged by the REMUS UUV’s horizontal and vertical sonar arrays (Horner et al.,
2009). Using this occupancy grid framework, each candidate trajectory’s risk of obstacle
collision is computed using the occupancy probabilities (a direct lookup operation) of the
grid cells it traverses. Trajectory optimization for OA entails minimizing the cumulative risk
of collision along the entire trajectory.

6. Path following
While the REMUS UUV and SeaFox USV are both commercial vehicles with proprietary
autopilots, both provide communications interfaces that allow experimental sensor and
computer payloads to override the primary autopilot via high-level commands. The REMUS
RECON interface, for example, closely resembles the augmented autopilot depicted in Figs.
2 and 3 (although direct actuator inputs are only available for propeller and cross-body
thrusters settings). For full overriding control, a payload module must periodically send
valid commands containing all of the following: i) desired depth or altitude, ii) desired
vehicle or propeller speed, and iii) desired heading, turn rate, or waypoint location. The
developed trajectory generator (described in Section 3) outputs reference trajectories as
parameterized expressions for each coordinate in a spatial curve plus a speed factor to use
while traversing that curve. Using these expressions as reference trajectories, the 3D path
following controller developed earlier (Kaminer et al., 2007) can compute turn rate and pitch
rate commands required to drive a vehicle onto (and along) the desired trajectory. The
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 85

RECON interface, however, does not accept pitch rate commands (for vehicle safety
reasons). Therefore, in order to use the aforementioned path following controller to track 3D
trajectories with the REMUS UUV, controller outputs must be partitioned into horizontal
(turn rate) and vertical (depth or altitude) commands as described in the following section
(obviously, the SeaFox USV only uses the turn rate commands).

6.1 Horizontal plane

Consider the 2D problem geometry depicted in Fig.17, which defines an inertial {I} frame,
Serret-Frenet {F} error frame and body-fixed reference frame {b}. The kinematic model of the
vehicle (2)-(3) reduces to

⎡ x(t )⎤ ⎡U 0 cosψ (t )⎤
⎢ y(t )⎥ = ⎢U sinψ (t ) ⎥ (33)
⎣ ⎦ ⎣ 0 ⎦
with dynamics described by

ψ =r (34)

XI

ψ F xF Reference
XI XI
Trajectory
{F }
T
pc
N qF ψ xb
{I } yF
YI qI {b}
yb
Fig. 17. Horizontal path-following kinematics

By construction, the local trajectory planner produces an analytic expression for each
component of the spatial trajectory as a function of virtual arc length, pc (τ ) . We can also
compute analytic expressions for p′c (τ ) and p′′c (τ ) , the first and second derivatives,
respectively, of the spatial trajectory. Using the relationships in Fig.17, the errors can be
expressed in the Serret-Frenet frame {F} as

⎡ xF ⎤
qF = ⎢ ⎥ = FI R(q I − pc ) (35)
⎣ yF ⎦
where FI R = [ T , N , B]T is a rotation matrix constructed from the tangent, normal, and
binormal vectors of the Serret-Frenet error frame {F}. The tangent vector is computed from
the expression for the trajectory’s first derivative as:
86 Autonomous Underwater Vehicles

p′c (τ )
T= (36)
p′c (τ )

For the 2D problem, the normal vector components can be computed directly from the
tangent vector components: N x = −Ty and N y = Tx . Additionally, the signed curvature of
the trajectory can be computed using the expressions for the trajectory's first and second
derivatives as:

′′ (τ )pcx
pcy ′ (τ ) − pcx
′′ (τ )pcy
′ (τ )
κ (τ ) = 3
(37)
p′c (τ )

Taking the time derivative of q F , we obtain the following state space representation for the
error kinematics (i.e. the position and heading of the vehicle’s body-fixed frame {b} relative
to the Serret-Frenet frame {F}, which follows the desired trajectory):

xF = −l(1 − κ yF ) + U 0 cosψ e
y F = −l(κ xF ) + U0 sinψ e (38)
ψ e = ψ − ψ F = uψ − ψ F = uψ − κ l

where l is the path length of the desired spatial curve and l describes the speed at which a
virtual target travels along this curve.
The goal is to drive the vehicle’s position error ( qF ) and heading error (Ψe) to zero. This will
drive the vehicle to the commanded trajectory location ( p c ) and align its velocity vector
with the trajectory’s tangent vector (T). The control signal uψ must now be chosen to
asymptotically drive the vehicle position and velocity vectors onto the commanded
trajectory. We choose the candidate Lyapunov function

V=
1 2
2
(
xF + yF2 + (ψ e − δψ )2 ) (39)

where δψ is a shaping function that controls the manner in which the vehicle approaches
the path

⎛ − yF ⎞
δψ = sin −1 ⎜⎜ ⎟⎟ (40)
⎝ yF + d ⎠
with d > 0 an arbitrary constant.
Using some algebra, we choose the following control laws to ensure that V < 0 :

l = K 1xF + U 0 cosψ e
sinψ e − sin δψ (41)
uψ = κ l + δψ + K 2 (ψ e − δψ ) − U0 yF
ψ e − δψ

In these expressions K 1 , K 2 , and d can be used as gains to tune the closed-loop

performance of the path following controller.
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 87

6.2 Vertical plane

Now consider the REMUS UUV manoeuvring in the vertical plane using altitude
commands. For survey operations, REMUS is typically programmed to follow a lawnmower
pattern at a constant altitude above the sea floor determined by the desired sidescan sonar
range. Since the ADCP/DVL sensor continuously measures vehicle altitude above the
bottom, this operating mode ensures safe operation over undulating bottoms with slopes of
up to 45 degrees (Healey, 2004). Obstacle avoidance manoeuvres are required to safely
negotiate steeper slopes, step-like terrain features (e.g. sand bars or coral heads), or objects
proud of the ocean floor. As described earlier, since the REMUS FLS is mounted in a fixed
orientation, it may detect new obstacles while executing a manoeuvre to avoid the current
obstacle threat. Periodic or detection-based replanning can handle these situations. This
scenario was illustrated conceptually in Fig.7.
When negotiating a step-up ridge or sand bar whose extent is not known due to sonar
occlusion at the time of detection, it may not be desirable to follow the planned vertical
trajectory to its completion. Between planning iterations, a simple yet safer approach is to
revert back to constant altitude control once the vehicle reaches a position directly above
the detected object boundaries. This condition can be checked using a 2.5D version of the
3D bounding box intersection test described above. Figure 18 illustrates a simulation
whereby the REMUS FLS detects the leading edge of a ridge at maximum sonar range.
Image processing algorithms compute range to the object (80m) as well as its width (W,
into the page) and its height above the seafloor (5.5m) but cannot determine the length of
the ridge since it is occluded by its own leading edge. Therefore, the obstacle detection
algorithm generates a 3D bounding box measuring W m wide x 1.0m long (assumed) x
5.5m high.
While the IDVD-method planner generates a vertical trajectory in NED coordinates,
in shallow water it is safer to operate the vehicle in an altitude control mode. Therefore,
the vertical coordinate trajectory is converted from a depth plan into an altitude
plan by assuming constant water depth over the planning horizon and using the
relationships

D = hnom + z(0) hplan (τ ) = D − zplan (τ )

(42)
Δhplan (τ ) = hplan (τ ) − hnom hcmd (τ ) = hnom + Δhplan (τ )

where D is the water depth computed at planner initialization time, hnom is the nominal
altitude set point, and hcmd (τ ) is the altitude command sent to the autopilot via the RECON
interface. The resulting altitude plan Δhplan (τ ) is shown in Fig.18 as a deviation from the
nominal mission segment altitude. As seen, sending this altitude command directly to the
vehicle autopilot would cause an undesirable jump in the altitude profile once the
ADCP/DVL sensor measures the vehicle’s true altitude above the ridge. Instead, switching
the altitude command to the nominal mission segment altitude once the vehicle reaches the
ridge will produce the desired altitude profile). Note that even though Fig.18 depicts a
sudden drop in altitude when the vehicle passes the ridge while commanding the nominal
mission segment altitude, in practice the vehicle dynamics will ensure that the UUV
executes a smooth transition back to its nominal survey altitude.
88 Autonomous Underwater Vehicles

Fig. 18. Simulation results for vertical OA using UUV altitude control mode

7. Computer simulations and sea trials

The proposed path-planning method can be tailored to a specific vehicle or operational
domain by modifying the performance index J to incorporate vehicle or actuator dynamics
(feasibility constraints) and mission objectives such as OA or underwater rendezvous. This
section presents simulated and in-water experimental results for four different applications
which use the proposed trajectory optimization framework for UMV guidance: i)
underwater docking of a UUV with a mobile underwater recovery system (MURS); ii)
optimal exploitation of a terrain-relative feature map to improve UUV self-localization
accuracy; iii) 2D or 3D OA in cluttered environments; and iv) specific USV implementations
for sonar-based OA in riverine operations.

7.1 Underwater recovery

The goal of underwater recovery (Yakimenko et al., 2008) is to be able to compute a
rendezvous trajectory from any point on the UUV holding pattern to any point on the
MURS holding pattern as shown in Fig.19 (note hereinafter that depth values are shown as
negative numbers).
While the stochastic simulation shown in Fig.19 employs circular race tracks, in practice the
MURS would establish a race track that allowed it to travel back and forth along two long
track legs (see Fig.20). These legs are needed to allow sufficient time to contact the UUV
(which is assumed to be in its holding pattern somewhere within communication range) and
allow it to transit from its holding pattern to the rendezvous point. The proposed sequence
of events is to have the MURS (at position 1 in Fig.20) signal the UUV (at position 2) and
command it to proceed to a rendezvous point by a certain time. The UUV computes the
trajectory required to comply with the command. If the commanded rendezvous is feasible,
the UUV sends an acknowledgement message. Otherwise (i.e. the request violated some
constraint) the UUV sends a denial message (stage A in Fig.20) and requests that the MURS
command a different rendezvous point or time. The final point of the trajectory is located in
the approximate location of the MURS docking station at a given time. Knowing the
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 89

geometry of the MURS allows the planner to construct a “keep out” zone corresponding to
the MURS propeller and aft control surfaces. The UUV rendezvous trajectory must avoid
this area. Once the rendezvous plan has been agreed upon and acknowledged, both the
UUV and the MURS proceed to position 3 for rendezvous (stage B). Finally, at position 4 the
recovery operation (stage C) is completed.

Fig. 19. Manifold of initial and final conditions

Fig. 20. Proposed rendezvous scenario

The simulated rendezvous scenario assumes three stages: communication (A), execution (B),
and recovery (C), respectively. From the trajectory generation standpoint we are primarily
concerned with optimizing the path that would bring the UUV from its current position
(point 2) to a certain rendezvous state (point 3) in the preset time Tr proposed by the MURS,
while obeying all possible real-life constraints and avoiding the MURS keep out zone.
Figures 21 and 22 present a computer simulation in which a MURS is moving due east at
1m/s (1.94kn) with the docking station at a depth of 15m. A UUV is located 800 meters
away. The MURS wishes to conduct a rendezvous operation Tr minutes later and sends the
corresponding information to the UUV. This information includes the proposed final
position x f , y f , z f rendezvous course, speed, and time. Figure 21 shows several generated
trajectories, which meet the desired objectives for this scenario and also avoid an obstacle
located along the desired path to MURS. These trajectories differ by the arrival time Tr .
During handshaking communications with the MURS, the UUV determines whether the
suggested Tr is feasible. Of the four trajectories shown, the trajectory generated for
90 Autonomous Underwater Vehicles

Tr = 450s happens to be infeasible (the constraints on controls are violated). The solution of
the minimum-time problem for this scenario yielded 488 seconds as the soonest possible
rendezvous time.
The other three trajectories shown in Fig.21 are feasible. That means that the boundary
conditions are met (by construction) and all constraints including OA are satisfied (via
optimization). As an example, Fig.22 shows the time histories for the yaw rate ψ c and flight
path angle γ c vehicle control parameters as well as the UUV’s speed as it followed the
trajectory for Tr = 600s .

Fig. 21. Examples of rendezvous trajectories

Fig. 22. Constrained vehicle parameters for Tr = 600s

Stochastic simulations of the manifolds shown in Fig.21 illustrate that a successful rendezvous
can take place in all cases as long as Tr is greater than a certain value. Furthermore, they show
that minimization of the performance index using the IDVD method ensures that a smooth,
realizable trajectory is calculated in just a few seconds, regardless of the initial guess.
Converting code to an executable file in lieu of using an interpretative programming language
reduces execution time down to a fraction of a second.
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 91

7.2 Feature-based navigation

In the last decade, several different UUVs have been developed to perform a variety of
underwater missions. Survey-class vehicles carry highly accurate navigational and sonar
payloads for mapping the ocean floor, but these payloads make such vehicles very
expensive. Vehicles which lack these payloads can perform many useful missions at a
fraction of the cost, but their performance will degrade over time from inaccurate self-
localization unless external navigation aids are available. Therefore, it is interesting to
consider collaborative operations via a team of vehicles for maximum utility at reasonable
cost. The NPS CAVR has been investigating one such concept of operations called feature-
based navigation. This technique allows vehicles equipped only with a GPS receiver and
low cost imaging sonar to exploit an accurate sonar map generated by a survey vehicle. This
map is comprised of terrain or bottom object features that have utility as future navigational
references. This sonar map is downloaded to the low-cost follow-on vehicles before launch.
Starting from an initial GPS position fix obtained at the surface, these vehicles then navigate
underwater by correlating current sonar imagery with the sonar features from the survey
vehicle’s map. The localization accuracy of vehicles performing feature-based navigation
can be improved by maximizing the number of times navigational references are detected
with the imaging sonar. The following simulation demonstrates how the IDVD trajectory
generation framework can be tailored to this application. By incorporating a simple
geometric model of an FLS having a range of 60m, 30-degree horizontal FOV and operating
at a nominal ping rate of 1Hz, a new performance index was designed to favour candidate
trajectories, which point the sonar toward navigational references in the a priori feature map.
For this example, we sought trajectories that could obtain at least three sonar images of each
feature in the map. Figure 23 shows results of a computer simulation in which the number
of times each target was imaged by the sonar has been annotated. The resulting trajectory is
feasible (i.e. satisfies turn rate constraints) and yields three or more sonar images of all but
two targets.

Fig. 23. Simulation results for a feature-based navigation application

92 Autonomous Underwater Vehicles

7.3 Obstacle avoidance in cluttered environments

Another application which benefits from the aforementioned trajectory generation
algorithm is real-time OA in a highly cluttered environment. Figure 24 illustrates simulated
trajectories for avoiding a field of point-like objects in the 2D horizontal plane (e.g. a kelp
forest) and in all three dimensions (e.g. a mine field). In both simulations, the performance
index was designed to minimize deviations from a predefined survey track line while
avoiding all randomly generated obstacles via a CPA calculation. Terminal boundary
conditions for the OA manoeuvre were chosen to ensure the UUV rejoined the desired track
line before reaching the next waypoint (i.e. the manoeuvre terminated at a position 95%
along the track segment). Initial boundary conditions were chosen to simulate a random
obstacle detection which triggers an avoidance manoeuvre after the UUV has completed
about 10% of the predefined track segment. For illustration purposes, Fig.24 includes several
candidate trajectories evaluated during the optimization process although the algorithm
ultimately converged to the trajectory depicted with a thicker (red) line (CPA distances to
each obstacle appear as dashed lines).
Figure 25 shows the results from an initial sea trial of 3D OA that took place in Monterey
Bay on 9 December 2008. This experiment tested periodic trajectory generation and
replanning on the REMUS UUV using a simulated obstacle map comprised of oriented
bounding boxes. As seen in Fig.25, initially the REMUS UUV follows a predefined track
segment (dash-dotted line) at 4 meters altitude. At some point the vehicle’s FLS simulator
”detects” an obstacle (i.e. the current REMUS position and orientation place the virtual
obstacle within the range and aperture limits of the FLS). This activates the OA mode, and
the planner generates an initial trajectory (green) from the current vehicle position to the
final waypoint. REMUS follows this trajectory until the next planning cycle (4 seconds later)
when the vehicle generates a new trajectory and continues this path planning-path
following cycle.

7.4 Obstacle avoidance in restricted waterways

The NPS CAVR in collaboration with Virginia Tech (VT) is developing technologies to
enable safe, autonomous navigation by USVs operating in unknown riverine environments.
This project involves both surface (laser) and subsurface (sonar) sensing for obstacle
detection, localization, and mapping as well as global-scale (wide area) path planning, local-
scale trajectory generation, and robust vehicle control. The developed approach includes a
hybrid receding horizon control framework that integrates a globally optimal path planner
with a local, near-optimal trajectory generator (Xu et al., 2009).
The VT global path planner uses a Fast Marching Method (Sethian, 1999) to compute the
optimal path between a start location and a desired goal location based on all available map
information. While resulting paths are globally optimal, they do not incorporate vehicle
dynamics and thus cannot be followed accurately by the USV autopilot. Moreover, since
level set calculations are computationally expensive, global plans are recomputed only
when necessary and thus do not always incorporate recently detected obstacles. Therefore, a
complimentary local path planner operating over a short time horizon is required to
incorporate current sensor information and generate feasible OA trajectories. The IDVD-
based trajectory generator described above is ideally suited for this purpose. VT has
developed a set of matching conditions which guarantee the asymptotic stability of this
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 93

a)

b)
Fig. 24. Simulated 2D (a) and 3D (b) near-optimal OA trajectories

Fig. 25. REMUS sea trial results demonstrating periodic planning and path following
94 Autonomous Underwater Vehicles

framework. When these matching conditions are satisfied, the sequence of local trajectories
will converge to the global path’s goal location. If the local trajectories no longer satisfy
these conditions (usually because the global path is no longer compatible with recently
detected obstacles), the global path is recomputed.
Simulation results demonstrate the need for local trajectories that incorporate vehicle
dynamics and real-time sensor data (Fig.26). For this simulation, an initial level set map was
computed using an occupancy grid created by masking land areas as occupied and water
areas as unoccupied in an aerial image of the Sacramento River operating area. Performing
gradient descent on the level set from the USV’s initial position produces an optimal path
shown in blue. To simulate local trajectory generation with a stale global plan, the initial level
set map was not updated during the entire simulation. Meanwhile, to simulate access to real-
time sensor data, the local planner was provided with a complete sonar map generated during
a previous SeaFox survey of the area. In Fig.26, this sonar map has been overlaid on the a priori

a)

b)
Fig. 26. Simulated local OA trajectories
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 95

map with red and green colour channels representing the probability that a cell is occupied
or unoccupied, respectively. Black pixels represent cells with unknown status. A short green
line segment depicts the USV’s orientation when the local planner is invoked, and the
resulting trajectory is shown in yellow. The first simulation (Fig.26a) shows a local trajectory
which deviates from the stale global plan to avoid a sand bar detected with sonar. In the
second simulation (Fig.26b) the USV is initially heading in a direction opposite from the
global path, but the local planner generates a dynamically feasible trajectory to turn around
and rejoin the global path later.
To track these local trajectories, the 2D controller described in Section 6.1 was implemented
on the SeaFox USV by mapping the controller’s turn rate commands into rudder commands
understood by the SeaFox autopilot. After validating the turn rate controller design during
sea trials on Monterey Bay, the direct method trajectory generator and closed-loop path
following controller were tested on the Pearl River in Mississippi on 22 May 2010 (Fig.27).
For this test, the local planner used a sonar map of the operating area to generate the
trajectory (the cyan line) from an initial orientation (depicted by the yellow arrow) to a

Fig. 27. Path-following controller test on the Pearl River

96 Autonomous Underwater Vehicles

desired goal point (depicted by a circle). The SeaFox USV then followed it almost precisely
(the magenta line). As seen from Fig.27 the trajectory generator was invoked at an arbitrary
location while the USV was performing a clockwise turn. Since the USV was commanded to
return to its start location upon completion of this manoeuvre, the magenta line includes a
portion of this return trajectory as well (otherwise, the actual USV track would be nearly
indistinguishable from the reference trajectory on this plot).

8. Conclusion
An onboard trajectory planner based on the Inverse Dynamics in the Virtual Domain direct
method presented in this chapter is an effective means of augmenting an unmanned
maritime vehicle’s autopilot with smooth, feasible trajectories and corresponding controls. It
also facilitates incorporation of sophisticated sensors such as forward-looking sonar for
deliberative and reactive obstacle avoidance. This approach has been implemented on both
unmanned undersea and surface vehicles and has demonstrated great potential. Beyond its
ability to compute near-optimal collision-free trajectories much faster than in real time, the
proposed approach supports the utilization of any practically-sound compound
performance index. This makes the developed control architecture quite universal, yet
simple to use in a variety of applied scenarios, as demonstrated in several simulations and
preliminary sea trials. This chapter presented results from only a few preliminary sea trials.
Future research will continue development of the suggested trajectory framework in
support of other tactical scenarios.

9. Acknowledgements
The authors wish to gratefully acknowledge the support of Doug Horner, Co-Director of the
CAVR and Principle Investigator for the REMUS UUV and SeaFox USV research programs
at NPS. In addition, Sean Kragelund would like to thank his CAVR colleagues Tad Masek
and Aurelio Monarrez. Mr. Masek’s outstanding software development work to implement
obstacle detection and mapping with forward looking sonar made possible the OA
applications described herein. Likewise, the tireless efforts of Mr. Monarrez to continually
upgrade, maintain, and operate CAVR vehicles in support of field experimentation have
made a lasting contribution to this Center.

10. References
Basset, G., Xu, Y. & Yakimenko, O. (2010). Computing short-time aircraft maneuvers using
direct methods,” Journal of Computer and Systems Sciences International, 49(3), 145-176
BlueView Technologies, Inc. (2011). 2D Imaging sonar webpage. Available from:
www.blueview.com/2d-Imaging-Sonar.html
Bourke, P. (1992). Intersection of a line and a sphere (or circle). Professional webpage.
Available from: http://paulbourke.net/geometry/sphereline
Elfes, A. (1989). Using occupancy grids for mobile robot perception and navigation.
Computer, 22(6), 46-57
Real-Time Optimal Guidance and Obstacle Avoidance for UMVs 97

Furukawa, T. (2006). Reactive obstacle avoidance for the REMUS underwater autonomous vehicle
using a forward looking sonar. MS Thesis, NPS, Monterey, CA, USA
Gadre, A., Kragelund, S., Masek, T., Stilwell, D., Woolsey, C. & Horner, D. (2009).
Subsurface and surface sensing for autonomous navigation in a riverine
environment. In: Proceedings of the Association of Unmanned Vehicle Systems
International (AUVSI) Unmanned Systems North America convention, Washington, DC,
USA
Healey, A. J. (2004). Obstacle avoidance while bottom following for the REMUS
autonomous underwater vehicle. In: Proceedings of the IFAC conference, Lisbon,
Portugal
Horner, D. & Yakimenko, O. (2007). Recent developments for an obstacle avoidance system
for a small AUV. In: Proceedings of the IFAC conference on Control Applications in
Marine Systems, Bol, Croatia
Horner, D., McChesney, N., Kragelund, S. & Masek, T. (2009). 3D reconstruction with an
AUV-mounted forward-looking sonar. In: Proceedings of the International symposium
on Unmanned Untethered Submersible Technology (UUST09), Durham, NH, USA
Hydroid, Inc. (2011). REMUS 100 webpage. Available from:
www.hydroidinc.com/remus100.html
Kaminer, I., Yakimenko, O., Dobrokhodov, V., Pascoal, A., Hovakimyan, N., Cao, C., Young,
A. & Patel, V. (2007). Coordinated path following for time-critical missions of
multiple UAVs via L1 adaptive output feedback controllers. In: Proceedings of the
AIAA Guidance, Navigation, and Control conference, Hilton Head, SC, USA
Kreuzer, J. (2006). 3D programming – weekly: Bounding boxes. Collision detection tutorial
webpage. Available from: www.3dkingdoms.com/weekly/weekly.php?a=21
Masek, T. (2008). Acoustic image mModels for navigation with forward-looking sonars. MS Thesis,
NPS, Monterey, CA, USA
Northwind Marine. (2011). SeaFox webPage. Available from:
www.northwindmarine.com/military-boats
Sethian, J. (1999). Fast marching method. SIAM Review, 41(2), 199-235
Xu, B., Kurdila, A. J. & Stilwell, D. J. (2009). A hybrid receding horizon control method for
path planning uncertain environments. In: Proceedings of the IEEE/RSJ International
conference on Intelligent Robots and Systems, St. Louis, MO, USA
Yakimenko, O. & Slegers, N. (2009). Optimal control for terminal guidance of autonomous
parafoils. In: Proceedings of the 20th AIAA Aerodynamic Decelerator Systems Technology
conference, Seattle, WA, USA
Yakimenko, O. (2000). Direct method for rapid prototyping of near optimal aircraft
trajectories. Journal of Guidance, Control, and Dynamics, 23(5), 865-875
Yakimenko, O. (2011). Engineering computations and modeling in MATLAB/Simulink. AIAA
Education Series, ISBN 978-1-60086-781-1, Arlington, VA, USA
Yakimenko, O. A. (2008). Real-time computation of spatial and flat obstacle avoidance
trajectories for AUVs. In: Proceedings of the 2nd IFAC workshop on Navigation, Guidance
and Control of Underwater Vehicles (NGCUV’08), Killaloe, Ireland
98 Autonomous Underwater Vehicles

Yakimenko, O.A., Horner, D.P. & Pratt, D.G. (2008). AUV rendezvous trajectories generation
for underwater recovery, In: Proceedings of the 16th Mediterranean conference on
Control and Automation, Corse, France
 0
5

Formation Guidance of AUVs Using

Decentralized Control Functions
Matko Barisic, Zoran Vukic and Nikola Miskovic
University of Zagreb, Faculty of Electrical Engineering and Computing
Croatia

1. Introduction
Autonomous Underwater Vehicles (AUVs) are the most complex type of unattended marine
systems, being mobile, with challenging dynamics and non-holonomic kinematics. They are
increasingly being recognized as a keystone technology for projecting human scientific and
economical interests into the deep Ocean (Papoulias et al., 1989). A recent report by Bildberg
(2009) delivers the verdict of several key researchers that the AUVs are rapidly moving
towards maturity.
The autonomy of AUVs is their key capability. They autonomously explore Ocean phenomena
relevant to human scientific and economic interests. Well engineered autonomous control
allows them to act robustly and predictably with regards to waves, currents, wind, sea-state
and numerous other disturbances and operational conditions in nature. As a consequence,
they are today being cast in the leading role in projecting human presence and human interests
in the Ocean, in an increasingly diverse gamut of topics:
• Physical oceanography (Plueddemann et al., 2008; Tuohy, 1994),
• Marine biology, conservationist biology, marine ecology management, biological
oceanography (Farrell et al., 2005; Pang, 2006; Pang et al., 2003),
• Geology, petrology, seismology, hydrography (for the benefit of e.g. the oil and gas
industry, maritime civil engineering etc.),
• Maritime and naval archaeology, submerged cultural heritage protection and
management,
• Marine traffic management, search and rescue, hazardous material and waste
management, emergencies and catastrophes management and first responding
(Carder et al., 2001; Pang, 2006)
• Maritime security, customs enforcement, border protection and defense (Allen et al., 1997;
2004; Clegg & Peterson, 2003; Curtin et al., 1993; Eisman, 2003; US Navy, 2004).
To increase the effectiveness, safety, availability, economics and applicability of AUVs to
these and other topics of interest, this chapter proposes a decentralized cooperative cross-layer
formation-control paradigm for entire groups of AUVs collaborating in exploration tasks.
The AUVs are assumed to navigate on a common “flight ceiling” by using robust altitude
controllers, based on altimeter echosounder measurements. The proposed virtual potential
framework allows for the 2D organization of individual trajectories on such a “flight ceiling”.
100
2 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

The goal is to provide decentralized consensus-building resulting in synoptical situational

awareness of, and coordinated manoeuvring in the navigated waterspace. The paradigm is
formally developed and tested in a hardware-in-the-loop simulation (HILS) setting, utilizing a
full-state hydrodynamical rigid-body dynamic model of a large, sea-capable, long-endurance
Ocean-going vehicle. Existence of realistic, technically feasible sensors measuring proxy
variables or directly the individual kinematic or dynamic states is also simulated, as is the
presence of realistic, non-stationary plant and measurement noise.

1.1 The cooperative paradigm

Since 1970s, robotics and control engineers have studied the cooperative paradigm.
Cooperative control is a set of complete, halting algorithms and machine-realized strategies
allowing multiple individual agents to complete a given task in a certain optimal way. This
optimality results from the agents’ leveraging each other’s resources (e.g. manoeuvring
abilities) to more effectively minimize some cost function that measures a “budget” of the
entire task, in comparison to what each agent would would be capable of on their own
(without the benefit of the group).
In the marine environment, such “social” leveraging is beneficial in several ways. Firstly,
deployment of more AUVs significantly reduces the time needed to survey a given theater
of operations. This has enormous economic repercussion in terms of conserved hours or
days of usually prohibitively expensive ship-time (for the vessel that is rendering operational
support to the AUV fleet). Secondly, deployment of a larger number of AUVs diversifies the
risk to operations. In a group scenario, loss of a (small) number of AUVs doesn’t necessarily
preclude the achievement of mission goals. Lastly, if each of the group AUVs are furnished
with adaptive-sampling algorithms, such as in the chemical plume-tracing applications
(Farrell et al., 2005; Pang, 2006; Pang et al., 2003), deployment of multiple vehicles guarantees
much faster convergence to the points of interest.
Cooperative control frameworks are split into centralized and decentralized strategies. A
centralized cooperative control system’s task is to determine the actions of each agent
based on a perfectly (or as near perfectly as possible) known full data-set of the problem,
which consists of the state vectors of every agent for which the problem is stated. The
centralized system instantiates a globally optimal solution based on the assessment of
momentary resource-disposition of the entire ensemble, as well as based on the total, if
possibly non-ideal knowledge of the environment. The state data are usually collected by
polling all agents through a communication network. After the polling cycle, the centralized
system communicates the low-level guidance commands back to individual agents. This
approach allows for the emergence of a global optimum in decision-making on grounds of
all obtainable information, but heavily depends on fault-intolerant, quality-assured, high-bandwidth
communication.
In a decentralized approach, such as we have chosen to present in this chapter, each agent
possesses imperfect state and perception data of every other agent and of the observable
portion of the environment, and locally decides its own course of action. The greatest issue
in decentralized cooperative control is the achievement of a consensus between separately
reasoning autonomous agents.

2. The virtual potentials framework

To address the issue of reactive formation guidance of a number of AUVs navigating in a
waterspace, a method based on virtual or artificial potentials is hereby proposed. The virtual
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 1013

potentials alleviate some of the most distinct problems encountered by competing reactive
formation guidance strategies, which are prone to the following problems:
• Reliance on the perfect knowledge of a map of the waterspace,
• Lack of reaction to the decentralized, agent-local process of accumulating or perfecting
knowledge of the environment on top of the initially imperfect situational awareness of
each individual agent,
• Trajectory planning that is sub-optimal, or optimal based on a hard-coded criterion, without
possibility of adjusting or restating that criterion at run-time, because the cost function is
implicit in the choice of mathematical tools (such as a distinct set of curve formulations
used for trajectories etc.).
Stemming from these considerations, we propose a scheme where each AUV in a 2D formation
imbedded in the “flight ceiling” plane as previously discussed maintains a local imperfect
map of the environment. Every possible map only ever consists of a finite number of
instantiations of any of the three types of features:
1. A way-point that is commanded for the entire formation, w ∈ R 2
2. Obstacles which need to be circumnavigated in a safe and efficient manner, (O i ), ∀i =
1 . . . n obs O i ⊂ R 2 ,
3. Vertices of the characteristic cell of the chosen formation geometry, covered in more detail
in sec. 2.3.3 and 3.
With this in mind, let the virtual potential be a real, single valued function P : R 2 → R,
mapping almost every attainable position of an AUV on the “flight ceiling” to a real. Let
P-s total differential exists almost wherever the function itself is defined. P can be said to live
on the subspace of the full-rank state-space of the AUVs, C = R 6 × SE 3 . The state-space
of the AUV is composed of the Euclidean 6-space R 6 spanned by the angular and linear
 T  T
velocities, { vT ω T } = { u v w | p q r } ≡ R 6 and a full 3D, 6DOF configuration-space
 T  T
{ xT ΘT } = { x y z | ϕ ϑ ψ } which possesses the topology of the Special Euclidean group
of rank 3, SE 3 . Function P therefore maps to a real scalar field over that same C .
Furthermore, this framework will be restricted to only those P that can be expressed in terms
of a sum of finitely many terms:
n
∃n ∈ N | PΣ = ∑ Pi (1)
i =1
Where Pi is of one of a small variety of considered function forms. Precisely, we restrict our
attention to three function forms with each one characteristic of each of the three mentioned
types of features (way-point, obstacle, vertices of formation cells).
The critical issue in the guidance problem at hand is Euclidean 2D distance (within the “flight
ceiling”) between pairs of AUVs in the formation, and each AUV and all obstacles. Therefore,
our attention is further restricted to only such { Pi } ⊂ L(C → R ) with L being the space of all
functions mapping C to R whose total differential exists almost wherever each of the functions
is defined on C , which can be represented as the composition Pi ≡ pi ◦ di , pi : R 0+ → R, and
di : C → R 0+ a Euclidean 2D metric across the “flight ceiling”. Consequently, Pi is completely
defined by the choice of pi (d), the isotropic potential contour generator. Choices and design of
pi (d)-s will be discussed in sec. 2.3.
102
4 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

With all of the above stated, a decentralized total control function f : Z → R 2 is then defined
as a sampling, repeated at sample times k ∈ Z0 , of the 2D vector field E : W i → R 2 over a
subspace W i ⊆ R 2 ⊂ C , the navigable waterspace:

∀x ∈ W i ⊆ R 2 ⊂ C , E (x) = −∇ PΣ (x) (2)

  ( ag) 
Where W i = R 2 \ i Oi ∪ j Oj contains all of R 2 to the exclusion of closed connected
subsets of R 2 that represent interiors of obstacles, {O i } and those that represent safety areas
around all the j-th AUVs (j = i) other than the i-th one considered. W i is an open, connected
subset of R 2 , the “flight ceiling”, inheriting its Euclidean-metric-generated topology and
always containing the way-point w. Sampling E at the specific xi (k) ∈ W i , the location of
the i-th AUV, results in f i (k), the total decentralized control function for the i-th AUV at time k
and location xi .

2.1 Passivity
The decentralized total control function f is used as the forcing signal of an idealized
dimensionless charged particle of unit mass, modeled by a holonomic 2D double integrator. If
any AUV were able to behave in this manner, the AUV would follow an ideal conservative
trajectory given by:
 t
( 0)
xi ( t ) = E [xi (τ )] dτ 2 + xi (3)
τ =0
This ideal conservative trajectory, while stable in the BIBO sense, is in general not
asymptotically stable, nor convergent by construction. The simplest case when this doesn’t
hold is when E (x) is an irrotational1 vector field whose norm is affine in the x − w 2D
Euclidean distance:
E (x) = e x − w + E0 ; e ∈ [0, ∞ ) (4)
And whose direction is always towards w:
id
∀x, E(x) · (x − w) = e x − w 2
+ E0 x − w (5)

In that case (3) can be regarded as a linear second or third order system with two of the poles
in ±i. Such a system exhibits borderline-stable oscillation – a hallmark of its conservativeness.
An example of such BIBO-stable non-convergent oscillation is given in figure 1.
Note that this analysis is irrespective of the initial condition ẋ0 as long as (4, 5) approximate
E (x) sufficiently well in some open ε-ball centered on w. However, AUVs are in general
not able to actuate as ideal holonomic 2D double integrators. The introduction of any finite
non-zero lag in the above discussion, which is sure to exist from first physical principles in a
real AUV, is sufficient to cause dissipation and as a consequence passivity and convergence to
w.

2.2 Local minima

In addition to the problem of passivity, the virtual potential approach suffers from the existence
of local minima. Without further constraints, the nature of E (W i ) so far discussed doesn’t
( j)
preclude a dense, connected, closed state-subspace Ci 0 ⊆ C, containing uncountably many
(l )
initial vectors { x0 | l ∈ R } of “related” trajectories (with the indexing by AUV denoted by
1 Whose rotor or curl operator is identically zero.
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 1035

Fig. 1. An example of an oscillatory trajectory due to the conservativeness of the virtual

potential system.

Fig. 2. Example of a local minimum occurring in virtual potential guidance in a 2D

waterspace.

i, and the enumeration of the distinct points of convergence other than the way-point by j
omitted for clarity) that do not converge to the way-point w or a finitely large orbit around
( j)
it, but rather to another point x∞ (or a finitely large orbit around it). Therefore, for each of
these uncountably many “nearby” trajectories (to be visualized as a “sheaf” of trajectories
emanating from a distinct, well defined neigbourhood in W i for some range of initial linear
(l )
and angular velocities) there exists a lower bound tl after which xi (t > tl ) − x∞ ≤ xi ( t >
tl ) − w almost always. The set {tl } is also dense and connected.
104
6 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

( j)
Furthermore, there is no prejudice as to the number of such Ci 0 -s, i.e. there exists CiΣ0 ⊆
 ( j)
C , CiΣ0 = j Ci 0 .
There may be multiple disjoint dense, connected, closed sets of initial
conditions of the trajectory of the i-th AUV which all terminate in the same, or distinct local
minima. The enumerator j may even come from R (i.e. there may be uncountably many distinct
local minima, perhaps arranged in dense, connected sets – like curves or areas in R 2 ).
An example of an occurrence of a local minimum is depicted in figure 2. In order to resolve
local minima, an intervention is required that will ensure that either one of the following
conditions is fulfilled:
1. The set CiΣ0 is empty by construction.
2. A halting P-complete algorithm is introduced that for every x0 ∈ CiΣ0 , triggering at t0 with
ε(t0 ) = sup x(t > t0 ) − x∞ | x0 characterizing a ε-ball centered on the particular x∞ and
containing all x(t > t0 ), to intervene in E (W i ) guaranteeing that this entire ball is outside
(a possibly existing) new CiΣ0 (with x0 ← x(t0 )).
Out of the two listed strategies for dealing with local minima, the authors have published
extensively on strategy 2 (Barisic et al., 2007a), (Barisic et al., 2007b). However, strategy 1
id
represents a much more robust and general approach. A method guaranteeing CiΣ0 = ∅ by
designing in rotors will be described in sec. 2.4.

2.3 Potential contour generators and decentralized control functions

As for the potential contour generators pi (d) : R 0+ → R, their definition follows from the
global goals of guidance for the formation of AUVs. Bearing those in mind, the potential
contour generators of each feature type, p( o,w,c ) (for obstacle, way-point and formation cell
vertex, accordingly) are specified below.

2.3.1 Obstacles
po

d [m]

(a) Graph of po ( d) : R + → R. (b) Graph of po ( d(x)) : R + → R ◦ R2 → R + .

Fig. 3. The potential contour generator of an obstacle, po (d(x)).

 +
A
po (d) = exp − 1; lim po (d) = 0; lim po (d) = ∞ (6)
d d→∞ d → 0+

∂ A+ A+ ∂ ∂
po (d) = − 2 exp ; lim po (d) = 0; lim po (d) = ∞ (7)
∂d d d d→∞ ∂d d → 0+ ∂d
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 1057

Where:
- po (d) : R + → R is the potential contour generator of obstacles, a strictly monotonously
decreasing smooth single-valued Lebesgue-integrable function mapping a non-negative real
to a real,
- A+ ∈ R + \ {0} is a positive real independent parameter dictating the scale of the acceleration
away from the obstacle.

2.3.2 Way-points

A− − −2
d ≤ d0 : p
2 d
2 id A0 if A0
pw (d) = ; d = = ; p = (8)
A− 2A−
0 0
d > d0 : A0− (d − d0 ) + p0 p p

A − 2
∴ pw (d, d > d0 ) = A0− d − c − ; lim pw (d) = ∞; lim pw (d) = 0 (9)
2A p d→∞ d → 0+

∂ ∂ ∂
pw (d) = max( A− −
p d, A0 ); lim pw (d) = A0− ; lim pw (d) = 0 (10)
∂d d → ∞ ∂d d → 0+ ∂d
Where:
- pw (d) : R + → R is the potential contour generator of way-points, a strictly monotonously
increasing smooth single-valued Lebesgue-integrable function mapping a non-negative real
to a real,
- A− +
p ∈ R \ {0} is a positive real independent parameter dictating the scale of acceleration
towards the way-point in the area of proportional attraction,
- A0− ∈ R + \ {0} is a positive real independent parameter dictating the constant acceleration
towards the way-point outside the area of proportional attraction,
- d p ∈ R + \ {0} is a positive real independent parameter dictating the radius of the open ball
centered on the way-point that constitutes the area of proportional acceleration.
10
9.5
9
8.5
8
7.5
7
6.5
6
5.5
5
4.5
pw

4
3.5
3
2.5
2
1.5
1 −
0.5
A 0
0
d 0
−0.5
−1
−1.5
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
d [m]
(a) Graph of pw ( d) : R + → R. (b) Graph of pw ( d(x)) : R + → R ◦ R2 → R + .

Fig. 4. The potential contour generator of a way-point, pw (d(x)).

106
8 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

2.3.3 Cell vertices

A good candidate potential contour generator of formation cell vertices, which behaves
similar to a function with local support, is the normal distribution curve, adjusted for
attractiveness (i.e. of inverted sign).

d2
pc (d) = − A− d
c c exp 1 − 2 ; lim pc (d) = 0; lim pw (d) = − A− c (11)
2dc d→∞ d → 0+

∂ A− d2 ∂ ∂
pc (d) = c d exp 1 − 2 ; pc (d) 0 = 0; lim pc (d) = 0 (12)
∂d dc 2dc ∂d d → ∞ ∂d
 
∂2 A−c d2 d2
p c ( d ) = 1 − exp 1 − (13)
∂d2 dc d2c 2d2c
∂2 !
∴ dmax = arg pc (d) = 0 (14)
∂d2
id
dmax = ± dc (15)
∂
pc (d) id = A−
c (16)
∂d dmax = d c

Where:
- pc (d) : R + → R is the potential contour generator of cell vertices of characteristic cells of
a formation, a strictly monotonously increasing smooth single-valued Lebesgue-integrable
function mapping a non-negative real to a real,
- A− +
c ∈ R \ {0} is a positive real independent parameter dictating the scale of acceleration
towards the cell vertex at the distance of maximum acceleration towards the vertex
(equivalent to the valuation of A−
c · N (± σ ) on a Gaussian normal distribution curve),
- dc ∈ R + \ {0} is a positive real independent parameter dictating the radius of a sphere
at which the inflection in the potential contour generator occurs, i.e. the distance at which
maximum acceleration towards the vertex occurs (taking the place of σ in (12), which is
analogous to a Gaussian normal distribution curve).
The potential of a square formation cell surrounding an agent that figures as an obstacle is
represented in figure 5.

2.3.4 Reformulation in terms of decentralized control functions

The monotonicity of (6, 8, 11) ensures that the direction of the gradient of the potential,
∇ P (x)/ ∇ P (x) ∈ SO2 , is always ±ni = (x − xi )/ x − xi . Therefore, since (1, 2) are linear,
(2) can be solved analytically for any finite sum of terms of the form specified by (6, 8, 11) up
−
to the values of the independent parameters ( A+ , A− −
p , A0 , A c , dc ). The procedure follows:

− ∇ PΣ (x) = −∇ ∑ Pi (x)
i
= ∑ (−∇ pi (di (x))) (17)
i
∂
= −∑ p [ d (x)] · ni (x) (18)
i
∂di (x) i i

Equation (18) can be summarized by designating the terms in (7, 10, 12) as a( o,w,c ) respectively.
( o,w,c ) ( o,w,c )
The terms ai · ni , can likewise be denoted ai , respectively, and represent the
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 1079

Fig. 5. The potential contour generator of a formation agent.

potential-based decentralized control functions due to the i-th feature.

obstacles vertices
(o) (w) (c)
− ∇ PΣ = ∑ a (x ) · n i (x) + a i
i   
( x ) · nw ( x ) +
 
∑ a (x ) · n i (x)
i  
(19)
i i
( o) a( w ) (x) (c)
ai (x) ai (x)
obstacles vertices
(o) (c)
= ∑ a i ( x ) + a( w ) ( x ) + ∑ ai ( x ) (20)
i i

 + 
obstacles
A+ A − w−x
= ∑ exp ni (x) + min[ A−
p d (x ), A c ]
i d i ( x ) 2 d i ( x ) w−x
vertices − 
Ac d (x )2
+ ∑ di (x) exp 1 − i 2 (21)
i
dc 2dc

2.4 Rotor modification

As mentioned in sec. 2.2, the virtual potential approach to guidance is extremely susceptible
to the appearance of local minima. A robust and simple approach is needed to assure local
minima avoidance.
In terms of the vector field introduced by (2), the analytical solution of which is presented in
id
(20 – 21), stable local minima occur due to the irrotationality of the field, rot E (x) = 0.
In order to avoid irrotationality, and thereby local minima, decentralized control functions
proposed in (7, 10, 12) are redesigned, adding a rotor component:
(s) (s) (r )
ai ←−−− ai
redef
ai ← ai + ai (22)

Where:
- ai is the redefined total decentralized control function due to the i-th feature (the dash will
hereafter be omitted),
(s)
- ai is the stator decentralized control function as introduced in the preceding section,
108
10 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

(r )
denoted with the superscript (s) to contrast it with the newly introduced ai ,
(r )
-ai is the rotor decentralized control function, all of which are continuous real 2D vector fields
over the Euclidean 2-space (mapping R 2 to itself) such that they Jacobians exist wherever each
of them is defined.
(r )
The introduction of ai establishes a non-zero rot(E ) by design, as follows:

rot E (x) = ∑ a i (x ) = 0
i
(s) (r )
= rot ∑ ai (x) + rot ∑ ai (x )
i i
  
id
=0
(r )
= rot ∑ ai ( x ) (23)
i

With respect to the way-point, its potential influence on an AUV in this framework must
not be prejudiced in terms of the direction of approach. If a decentralized control function
of a way-point were augmented with a rotor part, the direction of aw would deviate from
line-of-sight. The same is true of formation cell vertices. Therefore, the only non-zero rotor
decentralized control functions are those of obstacles. As a result, (23) can be further simplified
to:
obstacles,
w.p.,
vertices obstacles
(r ) (r )
rot E (x) = rot ∑ ai (x) = rot ∑ ai ( x ) (24)
i
An individual obstacle rotor decentralized control function is defined below:

∀i = enum(obstacles)
a(x) = ar (x)âr (x) (25)
⎛ ⎞
(r ) (r )
Ai Ai
a r (x) = exp ⎝ ⎠ (26)
d i (x )2 d i (x )
⎡ ⎤
1 0 0
âr (x) = ⎣ 0 1 0 ⎦ · (r i (x) × [ni (x) (0)]T ) (27)
(0 0 1)
 
−xi
r i (x) = w w−xi
(0) · [ni (x) (0)]T (28)
⎧  T
⎪
⎪
⎪r i = 1 :
⎨
ni (x) × vv − ( vv · ni )ni (0)
 T
r i (x) = 0 ≤ r < 1 : w− x i
( ) × [ni (x) (0)]T (29)
⎪
⎪ i w− x i
0
⎪
⎩
otherwise : 0

Where:
(r )
- Ai ∈ R \ {0} is a positive real independent parameter dictating the scale of acceleration
perpendicular to the direction of fastest flight from the obstacle,
(r )
- ai ∈ R + is the magnitude of the rotor decentralized control function,
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 109
11

(r )
- âi ∈ SO 2 is the direction of the rotor decentralized control function,
- xi ∈ R 2 is the center of the i-th obstacle,
- ni (x) ∈ SO 2 is the unit vector in the direction of fastest flight from the i-th obstacle,
(r ) id
- r i (x) is the unit rotor direction generator, such that âi (x) = ri × ni (x),
- v ∈ R 2 is the current true over-ground velocity of the AUV (including possible sideslip)
projected onto the “flight ceiling”.
The rotor decentralized control function and the total decentralized control function consisting
of the superposition of the rotor and stator parts, are displayed in figure 6.

w w

E [m] E [m]

(a) A 2D display of â(r) ∈ R2 (b) A 2D display of â ∈ R2

(r )
Fig. 6. Direction of the rotor decentralized control function ai and the two-term
(s) (r )
a i = ai + ai decentralized control function.

3. Potential framework of formations

The formation introduced by the proposed framework is the line graph occurring at the tile
interfaces of the square tessellation of R 2 , represented in figure 7. Due to a non-collocated nature
of AUV motion planning, an important feature of candidate tessellations is that they be periodic
and regular, which the square tessellation is.

Fig. 7. The square tiling of the plane.

Each AUV whose states are being estimated by the current, i-th AUV, meaning j-th AUV, j = i)
is considered to be a center of a formation cell. The function of the presented framework for
potential-based formation keeping is depicted in figure 8. In an unstructured motion of the
cooperative group, only a small number of cell vertices attached to j-th AUVs ∀ j = i, if any,
110
12 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

(a) Disordered arrangement (b) Formation arrangement

Fig. 8. The potential masking of agents out of and in formation.

are partially masked by nodes. The i-th AUV is attracted strongest to the closest cell vertex, in
line with how attractiveness of a node varies with distance expressed in (11). In the structured
case in 8.b), presenting an ideal, undisturbed, non-agitated and stationary formation, all the
j-th AUVs in formation are masking each the attractiveness (w.r.t. the i-th AUV) of the vertex
they already occupy. At the same time they reinforce the attractiveness of certain unoccupied
vertices at the perimeter of the formation. The vertices that attract the i-th AUV the strongest
thus become those that result in the most compact formation. Notice in figure 8.b) how certain
vertices are colored a deeper shade of blue than others, signifying the lowest potential.
The square formation cell is a cross figure appearing at the interstice of four squares in the
tessellation, comprised of the j-th AUV and the four cell vertices attached to it, in the sense
that their position is completely determined based on the i-th AUV’s local estimation of j-th
( i) ( i)
AUV’s position, (x̂ j ), as in figure 7. The cell vertices are uniquely determined by x̂ j and an
independent positive real scaling parameter f .

4. The platform – A large Aries-precursor AUV

The vehicle whose dynamic model will be used to demonstrate the developed virtual potential
framework is an early design of the NPS2 Aries autonomous underwater vehicle which was
resized during deliberations preceding actual fabrication and outfitting. The resulting, smaller
Aries vehicle has been used in multiple venues of research, most notably (An et al., 1997;
Marco & Healey, 2000; 2001). As Marco & Healey (2001) describe, the vehicle whose model
dynamics are used has the general body plan of the Aries, displayed in figure 9, albeit scaled
up. The body plan is that of a chamfered cuboid-shaped fuselage with the bow fined using a
nose-cone. The modeled Aries-precursor vehicle, the same as the Aries itself, as demonstrated
in the figure, combines the use of two stern-mounted main horizontal thrusters with a pair
of bow- and stern-mounted rudders (four hydrofoil surfaces in total, with dorsal and ventral
pairs mechanically coupled), and bow- and stern-mounted elevators.
Healey & Lienard (1993) have designed sliding mode controllers for the Aries-precursor
vehicle, considering it as a full-rank system with states x = [ vT ω T xT Θ] =
[ u v w| p q r | x y z| ϕ ϑ ψ ]T , relying on the actuators:

u (t) = [ δr (t) δs (t) n (t)]T (30)

2 Naval Postgraduate School, 700 Dyer Rd., Monterey, CA, USA.
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 111
13

Fig. 9. The Aries, demonstrating the body-plan and general type of the model dynamics.
Image from the public domain.

Where:
- δr (t) is the stern rudder deflection command in radians,
- δs (t) is the stern elevator planes’ command in radians,
- n (t) is the main propellers’ revolution rate in rad/s.

4.1 Model dynamics of the vehicle

The dynamics published by Healey and Lienard are used in the HILS3 in the ensuing sections,
and were developed on the grounds of hydrodynamic modelling theory (Abkowitz, 1969;
Gertler & Hagen, 1967), exploited to great effect by Boncal (1987). The equations for the six
degrees of freedom of full-state rigid-body dynamics for a cuboid-shaped object immersed in
a viscous fluid follow, with the parameters expressed in Table 1.

4.1.1 Surge

  ρ 
m u̇ − vr + wq − x G (q2 + r2 ) + y G ( pq − ṙ ) + z G ( pr + q̇) = L4 X pp p2 + Xqq q2 + Xrr r2
2
 ρ 2  
+ X pr pr + L Xu̇ u̇ + Xwq wq + Xvp vp + Xvr vr + uq Xqδs δs + Xqδb /2 δbp + Xqδb /2 δbs
2
 ρ   
+ Xrδr urδr + L2 Xvv v2 + Xww w2 + Xvδr uvδr + uw Xwδs δs + Xwδb /2 δbs + Xwδb /2 δbp
2
2" # ρ ρ 
+ u Xδs δs δs + Xδb δb /2 δb2 + Xδr δr δr2 − (W − B ) sin ϑ + L3 Xδs n uqδs (n ) + L2 Xwδs n uwδs
2
2 2
 ρ 2 2
+ Xδs δs n u δs (n ) + L u X pro p
2 2
(31)
2

3 Hardware-in-the-loop simulation.
112
14 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

4.1.2 Sway

  ρ 
m v̇ + ur − wp + x G ( pq + ṙ ) − y G ( p2 + r2 ) + z G (qr − ṗ ) = L4 Yṗ ṗ + Yṙ ṙ + Ypq pq
2
 ρ   ρ 
+Yqr qr + L3 Yv̇ v̇ + Yp up + Yr ur + Yvq vq + Ywp wp + Ywr wr + L2 Yv uv + Yvw vw
2 2

2  ρ xtail  2  v + xr
+Yδr u δr − Cdy h( x )(v + xr ) + Cdz b ( x )(w − xq )
2
dx + (W − B ) cos ϑ sin ϕ
2 xnose Uc f ( x )
(32)

4.1.3 Heave

  ρ 
m ẇ − uq + vp + x G ( pr − q̇ ) + y G (qr + ṗ ) − z G ( p2 + q2 ) = L4 Zq̇ q̇ + Z pp p2 + Z pr pr
2
 ρ   ρ  "
+ Zrr r2 + L3 Zẇ q̇ + Zq uq + Zvp vp + Zvr vr + L2 Zw uw + Zvv v2 + u2 Zδs δs + Zδb /2 δbs
2 2
# ρ  xnose   w − xq
+ Zδb /2 δbp + Cdy h( x )(v + xr )2 + Cdx b ( x )(w − xq )2 dx
2 xtail Uc f ( x )
ρ ρ  
+(W − B ) cos ϑ cos ϕ + L3 Zqn uq (n ) + L2 Zwn uw + Zδs n uδs (n ) (33)
2 2

4.1.4 Roll


Iy q̇ + ( Ix − Iz ) pr − Ixy (qr + ṗ) + Iyz ( pq − ṙ) + Ixz ( p2 − r2 ) + m x G (ẇ − uq + vp)
 ρ   ρ 
− z G (v̇ + ur − wp) + L5 K ṗ ṗ + Kṙ ṙ + K pq pq + Kqr qr + L4 Kv̇ v̇ + K p up + Kr ur
2 2
 ρ 3  
+ Kvq vq + Kwp wp + Kwr wr + L Kv uv + Kvw vw + u Kδb /2 δbp + Kδb /2 δbs
2
2
ρ
+(y G W − y B B ) cos ϑ cos ϕ − (z G W − z B B ) cos ϑ sin ϕ + L4 K pn up (n )
2
ρ 3 2
+ L u K pro p (34)
2

4.1.5 Pitch


Ix ṗ + ( Iz − Iy )qr + Ixy ( pr − q̇) − Iyz (q2 − r2 ) − Ixz ( pq + ṙ ) + m y G (ẇ − uq + vp)
 ρ   ρ 
− z G (u̇ − vr + wq ) + L5 Mq̇ q̇ + M pp p2 + M pr pr + Mrr r2 + L4 Mẇ ẇ + Mq uq
2   2 
 ρ 3
+ Mvp vp + Mvr vr + L Muw uw + Mvv v + u Mδs δs + Mδb /2 δbs + Mδb /2 δbp
2 2
2

ρ xnose   w + xq
− Cdy h( x )(v + xr )2 + Cdz b ( x )(w − xq )2 x dx − ( x G W − x B B ) ·
2 xtail Uc f ( x )
ρ ρ 
· cos ϑ cos ϕ − (z G W − z B B ) sin ϑ + L4 Mqn qn (n ) + L3 Mwn uw
2 2

+ Mδs n u2 δs (n ) (35)
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 113
15

4.1.6 Yaw


Iz ṙ + ( Iy − Ix ) pq − Ixy ( p2 − q2 ) − Iyz ( pr + q̇ ) + Ixz (qr − ṗ ) + m x G (v̇ + ur − wp)
 ρ   ρ 
− y G (u̇ − vr + wq ) + L5 Nṗ ṗ + Nṙ ṙ + Npq pq + Nqr qr + L4 Nv̇ ċ + Np up
2 2
 ρ 3  ρ  xnose 
+ Nr ur + Nvq vq + Nwp wp + Nwr wr + L Nv uv + Nvw vw + Nδr u2 δr − Cdy ·
2 2 xtail
 w + xq
· h( x )(v + xr )2 + Cdz b ( x )(w − xq )2 x dx + ( x G W − x B B ) cos ϑ sin ϕ + (y G W − y B B ) ·
Uc f ( x )
ρ
· sin ϑ + L3 u2 Npro p (36)
2

4.1.7 Substitution terms

$
Uc f ( x ) = (v + xr )2 + (w − xq )2 (37)
n
X pro p = Cd0 (η |η | − 1); η = 0.012 ; Cd0 = 0.00385 (38)
u
√
sign(n ) Ct + 1 − 1
(n ) = −1 + ·√ (39)
sign(u ) Ct1 + 1 − 1
L2 η | η | L2
Ct = 0.008 ; Ct1 = 0.008 (40)
2.0 2.0

4.2 Control
The Aries-precursor’s low-level control encompasses three separate, distinctly designed
decoupled control loops:
1. Forward speed control by the main propeller rate of revolution,
2. Heading control by the deflection of the stern rudder,
3. Combined control of the pitch and depth by the deflection of the stern elevator plates.
All of the controllers are sliding mode controllers, and the precise design procedure is
presented in (Healey & Lienard, 1993). In the interest of brevity, final controller forms will
be presented in the ensuing subsections.

4.2.1 Forward speed

The forward speed sliding mode controller is given in terms of a signed squared term for
the propeller revolution signal, with parameters (α, β) dependent on the nominal operational
parameters of the vehicle, and the coefficients presented in table 1:
 
− ũ(t)
n (t)| n (t)| = (αβ) 1
αu (t)| u (t)| + u̇ c (t) − ηu tanh (41)
φu
ρL2 Cd
α= ; Cd = 0.0034
2m + ρL3 Xu̇
n rad m
β = 0 ; n0 = 52.359 ; u0 = 1.832 (42)
u0 s s
114
16 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

W = 53.4 kN B = 53.4 kN L = 5.3 m Ix = 13587 Nms2

Ixy = −13.58 Nms2 Iyz = −13.58 Nms2 Ixz = −13.58 Nms2 Iy = 13587 Nms2
Ix = 2038 Nms2 xG = 0.0 m x B = 0.0 m yG = 0.0 m
y B = 0.0 m zG = 0.061 m z B = 0.0 m g = 9.81m/s2
ρ = 1000.0 kg/m2 m = 5454.54 kg
X p p = 7.0 · 10−3 Xqq = −1.5 · 10−2 Xrr = 4.0 · 10−3 X pr = 7.5 · 10−4
Xu̇ = −7.6 · 10−3 Xwq = −2.0 · 10−1 Xv p = −3.0 · 10−3 Xvr = 2.0 · 10−2
Xqδs = 2.5 · 10−2 Xqδb /2 = −1.3 · 10−3 Xrδr = −1.0 · 10−3 Xvv = 5.3 · 10−2
Xww = 1.7 · 10−1 Xvδr = 1.7 · 10−3 Xwδs = 4.6 · 10−2 Xwδb /2 = 0.5 · 10−2
Xδs δs = −1.0 · 10−2 Xδb δb /2 = −4.0 · 10−3 Xδr δr = −1.0 · 10−2 Xqδs n = 2.0 · 10−3
Xwδs n = 3.5 · 10−3 Xδs δs n = −1.6 · 10−3
Yṗ = 1.2 · 10−4 Yṙ = 1.2 · 10−3 Ypq = 4.0 · 10−3 Yqr = −6.5 · 10−3
Yv̇ = −5.5 · 10−2 Yp = 3.0 · 10−3 Yr = 3.0 · 10−2 Yvq = 2.4 · 10−2
Yw p = 2.3 · 10−1 Ywr = −1.9 · 10−2 Yv = −1.0 · 10−1 Yvw = 6.8 · 10−2
Yδr = 2.7 · 10−2
Zq̇ = −6.8 · 10−3 Z p p = 1.3 · 10−4 Z pr = 6.7 · 10−3 Zrr = −7.4 · 10−3
Zẇ = −2.4 · 10−1 Zq = −1.4 · 10−1 Zv p = −4.8 · 10−2 Zvr = 4.5 · 10−2
Zw = −3.0 · 10−1 Zvv = −6.8 · 10−2 Zδs = −7.3 · 10−2 Zδb /2 = −1.3 · 10−2
Zqn = −2.9 · 10−3 Zwn = −5.1 · 10−3 Zδs n = −1.0 · 10−2
K ṗ = −1.0 · 10−3 Kṙ = −3.4 · 10−5 K pq = −6.9 · 10−5 Kqr = 1.7 · 10−2
Kv̇ = 1.3 · 10−4 K p = −1.1 · 10−2 Kr = −8.4 · 10−4 Kvq = −5.1 · 10−3
Kw p = −1.3 · 10−4 Kwr = 1.4 · 10−2 Kv = 3.1 · 10−3 Kvw = −1.9 · 10−1
Kδb /2 = 0.0 K pn = −5.7 · 10−4 K prop = 0.0
Mq̇ = −1.7 · 10−2 M p p = 5.3 · 10−5 M pr = 5.0 · 10−3 Mrr = 2.9 · 10−3
Mẇ = −6.8 · 10−2 Muq = −6.8 · 10−2 Mv p = 1.2 · 10−3 Mvr = 1.7 · 10−2
Muw = 1.0 · 10−1 Mvv = −2.6 · 10−2 Mδs = −4.1 · 10−2 Mδb /2 = 3.5 · 10−3
Mqn = −1.6 · 10−3 Mwn = −2.9 · 10−3 Mδs n = −5.2 · 10−3
N ṗ = −3.4 · 10−5 Nṙ = −3.4 · 10−3 N pq = −2.1 · 10−2 Nqr = 2.7 · 10−3
Nv̇ = 1.2 · 10−3 N p = −8.4 · 10−4 Nr = −1.6 · 10−2 Nvq = −1.0 · 10−2
Nw p = −1.7 · 10−2 Nwr = 7.4 · 10−3 Nv = −7.4 · 10−3 Nvw = −2.7 · 10−2
Nδr = −1.3 · 10−2 N prop = 0.0

Table 1. Parameters of the Model Dynamics

It is apparent from the above that the propeller rate of revolution command comprises a
term that accelerates the vehicle in the desired measure (u̇ c (t)), overcomes the linear drag
(u (t)| u (t)|), and attenuates the perturbations due to disturbances and process noise (σ̇u (t)).

4.2.2 Heading
The sliding surface for the subset of states governing the vehicle’s heading is given below, in
(43). The resulting sliding mode controller is contained in (44).

σr = −0.074ṽ (t) + 0.816r̃ (t) + 0.573 ϕ̃ (t) (43)

0.074ṽ (r ) + 0.816r̃ (t) + 0.573 ϕ̃ (t)
δr = 0.033v(t) + 0.1112r (t) + 2.58 tanh
0.1
It should be noted that ṽ(r ) seems to imply the possibility of defining some vc (t) for the
vehicle to track. This is impractical. The Aries-precursor’s thrust allocation and kinematics,
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 115
17

nonholonomic in sway, would lead to severe degradation of this sliding mode controller’s
performance in its main objective – tracking the heading. Lienard (1990) provides a further
detailed discussion of this and similar sliding mode controllers.

4.2.3 Pitch and depth

The main objective of the third of the three controllers onboard the Aries-precursor HIL
simulator, that for the combination of pitch and depth, is to control depth. For a vehicle
with the holonomic constraints and kinematics of the model used here, this is only possible
by using the stern elevators δs to pitch the vehicle down and dive. Accordingly, the sliding
surface is designed in (44), and the controller in (45).

σz (t) = q̃ (t) + 0.520ϑ̃ (t) − 0.011z̃ (t) (44)

σz (t)
δs (t) = −5.143q (t) + 1.070ϑ (t) + 4 tanh
0.4
= −5.143q (t) + 1.070ϑ (t)
q̃ (t) + 0.520ϑ̃ (t) − 0.011z̃ (t)
+4 tanh (45)
0.4

5. Obstacle classification, state estimation and conditioning the control signals

In this section, the issues of obstacle classification will be addressed, giving the expressions
for (di , ni ) of every type of obstacle considered, which are functions prerequisite to obtaining
Pi -s through composition with one of (6, 8, 11). Also, full-state estimation of the AUV
(modeled after the NPS Aries-precursor vehicle described in the preceding section), x̂ =
[ û v̂ ŵ p̂ q̂ r̂ x̂ ŷ ẑ ϕ̂ ϑ̂ ψ̂ ]T will be explored. Realistic plant and measurement noise (ñ, ỹ ), which
can be expected when transposing this control system from HILS to a real application
will be discussed and a scheme for the generation of non-stationary stochastic noise given.
Finally, the section will address a scheme for conditioning / clamping the low-level control
signals to values and dynamic ranges realizable by the AUV with the Aries body-plan. The
conditioning adjusts the values in the low-level command vector c = [ ac u c rc ψc ]T to prevent
unfeasible commands which can cause saturation in the actuators and temporary break-down
of feedback.

5.1 Obstacle classification

The problem of classification in a 2D waterspace represented by R 2 is a well studied topic.
We have adopted an approach based on modeling real-world features after a sparse set of
geometrical primitives – circles, rectangles and ellipses.
In the ensuing expressions, {xint } will be used for the closed, connected set comprising the
interior of the obstacle being described. Ti shall be a homogeneous, isomorphic coordinate
transform from the global reference coordinate system to the coordinate system attached to
the obstacle, affixed to the centroid of the respective obstacle with a possible rotation by some
ψi if applicable.

5.1.1 Circles
Circles are the simplest convex obstacles to formulate mathematically. The distance and
normal vector to a circle defined by (xi ∈ R 2 , ri ∈ R + ), its center and radius respectively,
116
18 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

are given below:

% & & '
& &
di : R 2 \ xint : &xint − xi & < ri → R +

d i (x) = x − xi − r i ∈ R + (46)
% & & '
& &
ni : R 2 \ xint : &xint − xi & < ri → SO 2
x − xi
n i (x) = (47)
x − xi

Robust and fast techniques of classifying 2D point-clouds as circular features are very well
understood both in theory and control engineering practice. It is easy to find solid algorithms
applicable to hard-real time implementation. Good coverage of the theoretic and practical
aspects of the classification problem, solved by making use of the circular Hough transform
is given in (Haule & Malowany, 1989; Illingworth & Kittler, 1987; Maitre, 1986; Rizon et al.,
2007).

5.1.2 Rectangles
The functions for the distance and normal vector (di (x), ni (x)), of a point with respect to a
rectangle in Euclidean 2-space defined by (xi ∈ R 2 , ai , bi ∈ R + , ψi ∈ [− π, π )), the center of
the rectangle, the half-length and half-breadth and the angle of rotation of the rectangle’s long
side w.r.t. the global coordinate system, respectively, are given below:
&  −1 & (
& &
int & a i 0 int &
di : R \ x : &
2
· T i (x ) & < 1 → R +
& 0 bi &
∞
⎧
⎪
⎪ | ı̂ · T ( x )| < a : | ĵ · T ( x )| − b
⎪
⎨
i i i i

d i (x) = | ĵ · T ( x )| < b : | ı̂
& i · T ( x )| − a
T &
i i i (48)
⎪ &  &
⎪
⎪ & &
⎩ otherwise : & i |T ( x )| − a i bi
2 2 &
&  & (
& −1 &
& a 0 &
ni : R 2 \ xint : & i · Ti (xint )& < 1 → SO2
& 0 bi &
∞
⎧
⎪ ai −
⎪|ı̂ · Ti (x)| < 2 : Ti {sign [ ĵ · Ti (x)] ĵ}
1
⎪
⎨ bi −1
ni (x) = | ĵ · Ti (x)| < 2 : Ti {sign [ı̂ · Ti (x)] ı̂} T (49)
⎪
⎪
⎪otherwise :
⎩ Ti −1 Ti (x) − ai sign [ı̂ · Ti (x)] bi sign [ ĵ · Ti (x)]

There is a large amount of published work dedicated to the extraction of the features of
rectangles from sensed 2D point-clouds. Most of these rely on Hough space techniques
(Hough & Powell, 1960) and (Duda & Hart, 1972) to extract the features of distinct lines in an
image and determine whether intersections of detected lines are present in the image (He & Li,
2008; Jung & Schramm, 2004; Nguyen et al., 2009).

5.1.3 Ellipses
The method of solving for a distance of a point to an ellipse involves finding the roots of
the quartic (57). Therefore, it is challenging to find explicit analytical solutions, although some
options include Ferrari’s method (Stewart, 2003) or algebraic geometry (Faucette, 1996). A
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 117
19

computer-based control system can, however, employ a good, numerically stable algorithm to
obtain a precise enough solution. The rudimentary part of analytic geometry that formulates
the quartic to be solved is given below in (50 - 57).
The equation of the ellipse with the center in the origin and axes aligned with the axes of the
coordinate system is:
 x  2  y 2
+ =1 (50)
a b
 T
The locus of its solutions is the ellipse, {xe = xe ye }. The analysis proceeds by considering
 T
those x = x y ∈ R 2 for which x − xe is normal to the ellipse. The equation of such a normal
is:
xn (τ ) = kτ + xe (51)
Where τ ∈ R is an independent parameter, the degree of freedom along the line and k is the
direction vector of the line, given below:
 x 2  y 2  xe
e e y e T
k=∇ + −1 = a2 b2
(52)
a b
id
It follows that if τ = t = arg x, i.e. xn (t) = x. Then, the following manipulation can be made:
 T  T
x − xe y − ye = tx e ty e (53)
a2 b2

 T  2
T
xe ye = a2 x b y (54)
t + a2 t + b2

Substituting the right-hand side of (54) into (50), the quartic discussed is obtained as:
 2  2
ax by
+ =1 (55)
t + a2 t + b2

( t + b 2 )2 a2 x 2 + ( t + a2 )2 b 2 y2 = ( t + a2 )2 ( t + b 2 )2 (56)

( t + a2 )2 ( t + b 2 )2 − ( t + b 2 )2 a2 x 2 − ( t + a2 )2 b 2 y2 = 0 (57)

The greatest root of (57), t, allows for the calculation of (di (x), ni (x)) in (51, 54), as given
below:
 
a0
di : R \ x : x Ti
2 int T
T − 1 (x) < 1 → R +
b0 i
& &
& & & a 2 x b 2 y T &
di (x) = x − xe = &kt& = t & e e
& t + a2 t + b2 &
&

)
(t + b2 )2 a4 x2e + (t + a2 )2 b4 y2e
=t (58)
( t + a2 )2 ( t + b 2 )2
)
(t + b2 )2 a4 [ı̂Ti (x)]2 + (t + a2 )2 b4 [ ĵTi (x))]2
=t (59)
( t + a2 )2 ( t + b 2 )2
118
20 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

 
a0
ni : R 2 \ xint : xT Ti T −1 (x ) <1 → SO2
b0 i
 T
xe ye
( i) k t + a2 t + b 2 &
T i ni ( x ) = = & & T & (60)
k & xe ye &
& &
& t + a2 t + b 2 &
 T
ı̂ Ti (x) ĵTi (x)
T i ni ( x ) = & t + a2 t + b 2 & (61)
& T &
& ı̂Ti (x) ĵTi (x) &
& &
& t + a2 t + b 2 &
⎧  T ⎫
⎪
⎪ T ( ) ĵ T ( ) ⎪
⎪
⎪
⎪
ı̂ i x i x ⎪
⎪
⎨ ⎬
ni ( x ) = T i −1 & t + a 2 t + b 2
& (62)
⎪ &   T&
⎪
⎪ & ı̂Ti (x) ĵTi (x) &⎪ ⎪
⎪
⎩& &⎪⎪
& t + a2 t + b 2 & ⎭

With the advent of cheap solid-state perception sensors in service robotics and aerial
photography in the last decade, publication on fast and robust ellipse-fitting of 2D point
clouds has intensified – (Ahn et al., 1999; Jiang et al., 2007; Pilu et al., 1996).

5.2 State estimation

 T  T
The full state x = vT ω T xT ΘT = u v w| p q r | x y z| ϕ ϑ ψ of the AUV will be estimated
using the Scaled Unscented Transform Sigma-Point Kalman Filter (SP-UKF) introduced by
van der Merwe (2004).
The Extended Kalman Filter formulations that feature more prominently in marine control
engineering state-of-the-art are capable of estimating the states of nonlinear model dynamics
by taking into account only first-order statistics of the states (with possible addition of plant
/ process noise. EKFs use Jacobians of the nonlinear operator(s) evaluated at the current
state estimate. The Unscented Kalman Filters (UKF), in contrast, use the original non-linear
model dynamics to propagate samples – called sigma-points, which are characteristic of the
current estimate of the statistical distribution of the states, influenced by process and measurement
noise. The Kalman gain is evaluated based on the covariance of state hypotheses thus
propagated vs. the covariance of the samples characteristic of the current estimate of the
statistical distribution of the measurements. The Kalman gain will award a higher gain to
those measurements for which a significant correlation is discovered between the state and
measurement hypotheses and for which the covariance of the measurement hypotheses
themselves is relatively small. The algorithm is listed in table 2.

5.3 Measurement and process noise

The AUV is assumed to carry a 4-beam DVL4 which it can use to record the true 3D
 T
speed-over-ground measurement v = u v w . Furthermore, the AUV carries a 3-axial
 T
rate gyro package capable of measuring the body-fixed angular velocities ω = p q r . A

4 Doppler Velocity Logger.

Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 119
21

Table 2. The Scaled Unscented Transform Sigma-Point Kalman Filter Algorithm

0. Parameterization
0.1. Set α, the scaling parameter for the Scaled
Unscented Transform
0.2. Set β, the parameter of accentuation
of the central estimate
0.3. Set κ, the scaling parameter of the set of
sigma-point drawn from the underlying distribution
0.4. L = 12, the number of states
0.5. λ = α2%( L + κ ) − L '
0.6. wc = w(c0) w(c1) · · · w(c2L) ,
( 0) λ
wc = L+λ + ( 1 − α2 + β ) ,
(1...2L) (1...2L)
wc = wm = 1
2( L+ λ'
)
%
0.7. wm = w(m0) w(m1) · · · wm
(2L) ,

( 0)
λ
wm = L+ λ
1. Initialization
1.1. Set x̂(0|1) = x
- the initial a priori
 estimate 
1.2. Set P x (k) = E (x − x)(x − x)T
- the initial covariance matrix of the estimates
1.3. Set Rf - the process noise covariance
1.4. Set Rn - the measurement noise covariance
2. Iteration for k = 1 . . . ∞
2.1. Sigma-points and hypotheses of the states
2.1.1. X − (-k) = {x− } = . . /
x ( k | k − 1) x ( k | k − 1) + γ P x ( k ) x ( k | k − 1) − γ P x ( k )
2.2. Time-update - /
2.2.1. X − ∗ (k) = {x− ∗ } = F (X − (k))
( i)
2.2.2. x̂− (k| k − 1) = ∑2L w x− ∗ ( i)
 i =0 m  T
2L ( i )
2.2.3. P − x = ∑ i wc X − ∗ ( i) − x̂− (k| k − 1) X − ∗ ( i) − x̂− (k| k − 1)
2.2.4. Re-draw the hypotheses taking into
account - the process √ noise covariance√ /
X (k| k − 1) = x̂− x̂− + γ Rv x̂− − γ Rv
2.2.5. Y (k| k − 1) = H (X )
( i) ( i)
2.2.6. ŷ − = ∑2L i =0 w m Y
2.3. Measurement update
  T
( i) ( i ) − ŷ − ( i ) − ŷ −
2.3.1. P y = ∑2L i =0 w c Y Y
  T
( i)
2.3.2. P xy = ∑2L i =0 w c X ( i) − x̂− Y ( i) − ŷ −
2.3.3. K(k) = P xy P − y
1
" #
2.3.4. x̂(k| k) = x̂ (k| k − 1) + K(k) y − ŷ −
−

2.3.5. P x (k) = P − x − K(k)P y K(k)

T
120
22 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

 T
low-grade commercial USBL5 system is assumed to provide estimates of x y . A fusion
of the USBL estimate and the pressure gauge prior to the SP-UKF entry point is assumed to
provide a relatively good-quality depth reading of z. A 3-axial middle-market strap-down
 T
AHRS6 is assumed to provide the Tait-Bryan angle readings, Θ = ϕ ϑ ψ .
In the proposed HILS framework, measurement noises should mimic the actual experience
during AUV fieldwork operations. Therefore, a noise generator which can produce
non-stationary, varying noises is required. It is intended that these noises include errors in the
sensor readings whose sources cannot be simply identified by recourse to first-order statistics,
and which therefore cannot be easily calibrated (de-biased) for. Additionally, we wish to be
able to generate sporadic irrecoverable faults i.e. events during which a sensor reading cannot
be relied on in any meaningful way.
For this reason, we propose the use of a bank of Gaussian Markov models – GMMs, for
generating the additive measurement noise. Markov models are stochastic state-machines
whose state-switching is governed by random number generators. GMMs ultimately output
a normally distributed random number with the statistics dependent on the current state.
Means and standard deviations (μ i , σi ) of each state i are designed into the GMM. In this
chapter, a bank of 12n AUV Gaussian Markov models, one for each state of each of the n AUV
AUVs is used. All of the GMM states contain separate univariate rate-limited white noise
generators parameterized by (μ i , σi , n i ), where n i is the rate limit of the additive measurement
noise in the i-th channel.
Relying on the MATLAB normally distributed random number generator invoked by the
randn command, each state generates a number according to:

ỹ−
i ( k ) = μ i + σi · randn (63)
 n 
− −
ỹi (k) = sign(ỹ (k) − ỹ(k − 1)) · min ỹ (k) − ỹ (k − 1) , i (64)
T
Where T is the sampling time.
To optimize between a realistic nature of the measurement noises and HILS
complexity, each of the Markov models in the employed bank contains 6 states,
{nominal, +reliable, −reliable + unreliable, − unreliable, f ault}. The 6-state Gaussian
Markov models are initialized by a 6-tuple M = ((μ1 , σ1 , n1 ), · · · , (μ6 , σ6 , n6 )) and a
6 × 6 transition matrix T = [ tij ] with tij being a priori probabilities of switching from state
i to state j. The actual parameters used in the HILS simulation are presented in (65 – 74).
Before adding them to idealized state measurements, the noise channels are mixed together
(correlated) as y ← S y y using the matrix S y in (75), to mimic the physics of the relevant
sensors’ interdependence of measurements. Notice that S y has a pronounced block-diagonal
structure, indicative of the fact that the mentioned sensors (DVL, USBL, AHRS, gyro-compass
and rate gyros) output several state measurements each. The correlation between the states
measured by a single instrument is more pronounced than the one between measurements of
mutually dislocated sensors operating along different physical principles.
⎧
⎪
⎪ State nominal : ( μ1 , σ1 , n1 ) = (0, 0.06, 0.03815)
⎪
⎨State ± reliable : ( μ1 , σ1 , n1 ) = (±0.09, 0.11, 0.05)
Myv = Myu = Myv = Myw = (65)
⎪
⎪ State ± unreliable : ( μ1 , σ1 , n1 ) = (±0.2981, 0.24, 0.09)
⎪
⎩
State f aulty : ( μ1 , σ1 , n1 ) = (NaN, NaN, NaN)
5 Ultra-short baseline hydroacoustic localization.
6 Attitude and heading reference system.
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 121
23

⎡ ⎤
0.7542 0.1000 0.1000 0.0208 0.0208 0.0042
⎢0.4739 0.3365 0.1043 0.0379 0.0284 0.0190⎥
⎢ ⎥
⎢0.4739 0.1043 0.3365 0.0284 0.0379 0.0190⎥
Ty v = Ty u = Ty v = Ty w =⎢ ⎥ (66)
⎢0.3825 0.1749 0.1749 0.0984 0.0984 0.0710⎥
⎣0.3825 0.1749 0.1749 0.0984 0.0984 0.0710⎦
0.0270 0.1622 0.1622 0.2703 0.2703 0.1081
⎧ " π #
⎪
⎪ State nominal : ( μ1 , σ1 , n1 ) = 0, 85 , π
⎪
⎨State ± reliable : " π π227.608π #
( μ1 , σ1 , n1 ) = ± 72 , 60 , 144.201
M yω = M y p = M y q = M y r = " π π π
# (67)
⎪State ± unreliable :
⎪ ( μ1 , σ1 , n1 ) = ± 21.5 , 18.8 , 64.454
⎪
⎩
State f aulty : ( μ1 , σ1 , n1 ) = (NaN, NaN, NaN)
⎡ ⎤
0.5928 0.1596 0.1596 0.0423 0.0423 0.0033
⎢0.4978 0.2489 0.1511 0.0356 0.0533 0.0133 ⎥
⎢ ⎥
⎢0.4978 0.1511 0.2489 0.0533 0.0356 0.0133 ⎥
T yω = T y p = T y q = T y r =⎢
⎢0.5234
⎥ (68)
⎢ 0.1963 0.0561 0.0935 0.0748 0.0561 ⎥
⎥
⎣0.5234 0.0561 0.1963 0.0748 0.0935 0.0561 ⎦
0.0588 0.2941 0.2941 0.1176 0.1176 0.1176
⎧
⎪
⎪ State nominal : ( μ1 , σ1 , n1 ) = (0, 1.0, 0.012)
⎪
⎨State ± reliable : ( μ1 , σ1 , n1 ) = (±1.3, 1.5, 0.06)
My xy = My x = My y = (69)
⎪
⎪ State ± unreliable : ( μ1 , σ1 , n1 ) = (±3.85, 4.0, 1.28)
⎪
⎩
State f aulty : ( μ1 , σ1 , n1 ) = (NaN, NaN, NaN)
⎡ ⎤
0.4809 0.1967 0.1967 0.0601 0.0601 0.0055
⎢0.4160 0.3200 0.1440 0.0800 0.0320 0.0080 ⎥
⎢ ⎥
⎢0.4160 0.1440 0.3200 0.0320 0.0800 0.0080 ⎥
Ty xy = Ty x = Tyy =⎢
⎢0.3689
⎥ (70)
⎢ 0.2136 0.1359 0.1942 0.0777 0.0097 ⎥
⎥
⎣0.3689 0.1359 0.2136 0.0777 0.1942 0.0097 ⎦
0.0102 0.2959 0.2959 0.1837 0.1837 0.0306
⎧
⎪
⎪ State nominal : ( μ1 , σ1 , n1 ) = (0, 0.08, 0.012)
⎪
⎨State ± reliable : ( μ1 , σ1 , n1 ) = (±0.11, 0.1208, 0.06)
Myz = (71)
⎪State ± unreliable :
⎪ ( μ1 , σ1 , n1 ) = (±0.55, 0.71, 1.28)
⎪
⎩
State f aulty : ( μ1 , σ1 , n1 ) = (NaN, NaN, NaN)
⎡ ⎤
0.5198 0.1762 0.1762 0.0617 0.0617 0.0044
⎢0.4020 0.4020 0.1106 0.0503 0.0302 0.0050⎥
⎢ ⎥
⎢0.4020 0.1106 0.4020 0.0302 0.0503 0.0050⎥
Tyz =⎢
⎢0.3704
⎥ (72)
⎢ 0.2667 0.1481 0.1481 0.0593 0.0074⎥⎥
⎣0.3704 0.1481 0.2667 0.0593 0.1481 0.0074⎦
0.0667 0.2000 0.2000 0.2000 0.2000 0.1333
⎧ " π
#
⎪
⎪ State nominal : ( μ1 , σ1 , n1 ) = 0, 220 , π
⎪
⎨State ± reliable : " π 98.05π
#
( μ1 , σ1 , n1 ) = ± 192 , 176 , π
MyΘ = My ϕ = Myϑ = Myψ = " π π π 42.60 # (73)
⎪
⎪ State ± unreliable : ( μ1 , σ1 , n1 ) = ± 60 , 42 , 10
⎪
⎩
State f aulty : ( μ1 , σ1 , n1 ) = (NaN, NaN, NaN)
⎡ ⎤
0.4686 0.2301 0.2301 0.0335 0.0335 0.0042
⎢0.4014 0.2721 0.1769 0.1020 0.0340 0.0136 ⎥
⎢ ⎥
⎢0.4014 0.1769 0.2721 0.0340 0.1020 0.0136 ⎥
TyΘ = Ty ϕ = Tyϑ = Tyψ =⎢
⎢0.3982
⎥ (74)
⎢ 0.1403 0.0995 0.1719 0.0995 0.0905 ⎥
⎥
⎣0.3982 0.0995 0.1403 0.0995 0.1719 0.0905 ⎦
0.0526 0.1579 0.1579 0.2105 0.2105 0.2105
122
24 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

⎡ 1 1.0 · 10− 4 1.0 · 10− 4 3.0 · 10− 5 3.0 · 10− 5 3.0 · 10− 5 0 0 0 5.0 · 10− 7 5.0 · 10− 7 5.0 · 10− 7
⎤
1.0 · 10− 4 1 1.0 · 10− 4 3.0 · 10− 5 3.0 · 10− 5 3.0 · 10− 5 0 0 0 5.0 · 10− 7 5.0 · 10− 7 5.0 · 10− 7
⎢ 1.0 · 10− 4 1.0 · 10− 4 3.0 · 10− 5 3.0 · 10− 5 3.0 · 10− 5 5.0 · 10− 7 5.0 · 10− 7 5.0 · 10− 7 ⎥
⎢ 3.0 · 10− 5 3.0 · 10− 5
1
3.0 · 10− 5 6.0 · 10− 4 6.0 · 10− 4
0 0 0
⎥
⎢ 1 0 0 0 0 0 0 ⎥
⎢ 3.0 · 10− 5 3.0 · 10− 5 3.0 · 10− 5 6.0 · 10− 4 1 6.0 · 10− 4 0 0 0 0 0 0 ⎥
⎢ 3.0 · 10− 5 3.0 · 10− 5 3.0 · 10− 5 6.0 · 10− 4 6.0 · 10− 4 ⎥
Sy = ⎢ 0 1 0 0 0
3.0 · 10− 4 4.0 · 10− 6
0 0
2.0 · 10− 5 2.0 · 10− 5
0
2.0 · 10− 5
⎥ (75)
⎢ 0 0 0 0 0 1
⎥
⎢ 0 0 0 0 0 0 3.0 · 10− 4 1 4.0 · 10− 6 2.0 · 10− 5 2.0 · 10− 5 2.0 · 10− 5 ⎥
⎢ 0 0 0 0 0 0 4.0 · 10− 6 2.0 · 10− 5 1 0 3.0 · 10− 5 3.0 · 10− 5 ⎥
⎣ 5.0 · 10−7 5.0 · 10− 7 5.0 · 10− 7 0 0 0 2.0 · 10− 5 4.0 · 10− 6 1 0 0 0 ⎦
5.0 · 10− 7 5.0 · 10− 7 5.0 · 10− 7 0 0 0 2.0 · 10− 5 2.0 · 10− 5 0 3.0 · 10− 5 1 3.0 · 10− 5
5.0 · 10− 7 5.0 · 10− 7 5.0 · 10− 7 0 0 0 2.0 · 10− 5 2.0 · 10 − 5 0 3.0 · 10 − 5 3.0 · 10− 5 1

5.3.1 Sensor fault simulation

The last mentioned state, “fault” doesn’t generate random additive measurement noises, but
outputs NaN7 values, which are ignored by the SP-UKF. On detecting the NaN value in a
measurement channel, the SP-UKF sets the corresponding, i-th row of the column vector of
measurements yi to the value of ŷi . This results in the corresponding rows and columns of the
P y covariance matrix being zero, an update to the corresponding row of the estimate column
vector x̂(k| k) being omitted, and the corresponding rows and columns of P x (k) growing
rather than falling. The latter signifies a decrease in the trustworthiness of the estimate, and
compromises the stability of the SP-UKF should a string of faulty readings continue overlong.

5.3.2 Outlier rejection

In order for the SP-UKF to remain stable and deliver trustworthy state estimates to the
feedback of the relevant controllers, outlier measurements are rejected. Rejection of an outlier
is dealt with identically to a faulty measurement, the appropriate row of the measurement
column vector y being over-written with NaN as if there were a sensor fault occurring in the
outlier measurement channel.
Outlier rows of y are considered those for which any values, inspected column-wise, fulfill:
 
( i,j )
R y ( k ) = y ( k )T · y ( k ) = r y (76)
%  '
( i,j )
iout = argi ∃ j, ry > 16 · P x (k) (77)
redef
y [{iout }] = NaN (78)

Where P x (k) is the covariance matrix of the estimate of the full-state vector rendered by the
SP-UKF.

5.3.3 Process noise

The additive process noise is assumed to be multivariate rate-limited white noise without bias.
The used covariance matrix is given in (79) and the rate limits in a vector in (80).
⎡ ⎤
0.2 0.01 0.01 0 0.005 0.005
⎢ 0.01 0.1 0.0275 0.01 0.003 0.035 ⎥
⎢ ⎥
( true ) ⎢ 0.01 0.0275 0.1 0.01 0.003 0 ⎥
Rv =⎢
⎢ 0
⎥ (79)
⎢ 0.01 0.01 0.0011 0.0001 0.00015 ⎥ ⎥
⎣0.005 0.003 0.03 0.0001 0.002 0 ⎦
0.005 0.035 0 0.00015 0 0.0022
( true )  π π π T

nv = 0.2 0.2 0.2 36 36 36 (80)
7 Not a Number.
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 123
25

5.4 Control signals conditioning

The solution of f i (k) ← E (x̂ AUV (k)), obtained by evaluating (2), or more precisely (29), at
the estimate x̂ AUV (k) rendered by the SP-UKF, will be used for the formation of commands
accepted by the forward speed and heading controllers in (41, 44), c = [ u̇ c u c ψ̇c ψc ].
The task of the low-level control system of the AUV is to try to recreate a motion that would
result from applying f (k) to an unconstrained point unit mass, i.e. a holonomic 2D double
integrator model, up to the thrust allocation and kinematic and dynamic constraints of the
actual vehicle. In the following equations, the sampling with time T of the integration over
the R 2 space is assumed in the form of the Euler backwards formula. So naively:
$
u (k) = u (k − 1)2 + T 2 f 2 + 2Tu f (81)

1
u̇(k) = (u (k) − u (k − 1))
T
1$ u ( k − 1)
= u (k − 1)2 + T 2 f 2 + 2T f u (k − 1) − (82)
T T
2 3
1 T f⊥
ψ̇(k) = arctan (83)
T T f + u ( k − 1)
4 2 3 5
1 1 T f⊥
ψ̈(k) = arctan − ψ̇(k − 1) (84)
T T T f + u ( k − 1)

Where:
id  T
- f = f is the norm of the total controlling force, admitting decomposition into f f ⊥ ,
the components parallel and perpendicular to the direction of heading of the AUV (given by
ψ (k − 1), notwithstanding possible sideslip resulting from x2 = v2 = v = 0).
At this point, it is assumed that an AUV has a specified performance envelope of (u, u̇, ψ̇, ψ̈ ).
With these as given independent parameters, the manipulation of (81 - 84) results in the
constraints that dictate the admissible ranges to which f needs to be clamped to avoid forcing
the low-level controllers of the AUV beyond their normal operating range.

5.4.1 The constraint inequalities

The locus of solutions of (81) for f in easier to visualize in an AUV-fixed coordinate system
(one with the origin in xi , with the x-axis aligned with u (k − 1)∠ψ (k − 1)). In it, the admissible
solution locus is a disc offset along the x-axis by u (k − 1)/T, given by the implicit expression:
& T &
& & u
& f + u ( k −1) f &
& T ⊥ &≤ T (85)

The locus of solutions of (82) is, similarly to the preceding case, an offset annulus concentric to
the disc described by (85):
& T &
u ( k − 1) & &
− u̇ ≤ & f + u ( k −1)
f & ≤ u (k − 1) + u̇ (86)
T & T ⊥ & T
124
26 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

The locus of solutions of (83) is an angular sector of an infinite disk (a 1-cone) concentric to the
previous two loci, expressed in terms of:
⎛ ⎞
a ⊥
arctan ⎝ u ( k −1)
⎠ ≤ T ψ̇ (87)
a + T

The locus of solutions of (84) is likewise an angular sector of an infinite disk (a 1-cone) concentric
to all the other loci, which satisfies the inequality:
⎛ ⎞
T2 f ⊥
T ψ̇ (k − 1) − ψ̈ ≤ arctan ⎝ u ( k −1)
⎠
2 f + T
T2
≤ T ψ̇ (k − 1) + ψ̈ (88)
2
A solution for f (k) is legal if it meets all of the criteria stated in (85–88), i.e. if it belongs to a
subset of R 2 shaped as an annular sector.

5.4.2 Clamping the total controlling force

If (85 – 88) are not fulfilled, a non-linear procedure for the clamping f to an admissible range
is employed. Consequently, the low-level controllers’ operating point(s) remain within a
quasi-linear vicinity of the sliding surfaces. The procedure is pseudo-coded in Algorithm
3. After clamping by the presented algorithm, (81 – 84) are used to form the commands
 T
c = ac u c rc ψc for the low-level controllers 41, 44).

6. Simulation results
Combining the presented virtual potentials framework and the HILS presented in the
previous chapters, a full simulation is presented for a group of 4 simulated Aries-precursor
AUVs cruising in formation.

6.1 Simulation 1
The first simulation presents a cruise in formation down an unobstructed channel in between
two obstacles towards the way-point. Figure 10 presents the actual paths traveled by the
AUVs. Figure 11 presents the speeds of all four vehicles. The initial dips in the path occur due
to the non-holonomic nature of the vehicles’ kinematics, due to which they cannot initialize
the manoeuvres from zero starting speed that would preserve the initial formation perfectly
and still commence navigation to the way-point. This is especially exacerbated by the fact
that at near-zero speeds, the control surfaces (δr , δs ) are terribly ineffective. The final dips in
the area of the paths around the way-point occur after the AUVs have parked in the stable
formation configuration. Near the waypoint and at low speeds, the drift in the state estimates
is accentuated by a lack of passive stability provided by AUVs’ streamlining at higher speeds.
This, in hand with non-holonomic kinematics, causes the vehicles to momentarily break
formation. It is only after accumulating enough speed that vehicles can turn within a small
enough radius to re-establish the formation. Dips in the path correspond to the dips in
the speed graphs for the vehicles, as their commanded speed shoots up again in order to
re-establish the formation.
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 125
27

450

400

350

300

x [m] 250

200

150

100

50

0
0 50 100 150 200 250 300 350 400 450
y [m]
Fig. 10. Paths of the 4 HILS models of AUVs based on the precursor to the NPS Aries vehicle
cruising in an uncluttered environment.

u(t) [kts]

0 100 200 300 400 500 600 700 800 900

u(t) [kts]

0 100 200 300 400 500 600 700 800 900

u(t) [kts]

0 100 200 300 400 500 600 700 800 900

u(t) [kts]

0 100 200 300 400 500 600 700 800 900

Fig. 11. Speeds of the 4 HILS models of AUVs based on the precursor to the NPS Aries
vehicle cruising in an uncluttered environment.
126
28 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

if not (85) or (86)

if f > 0
— this indicates that the required dominant behaviour produced
by the virtual potentials framework is to accelerate the AUV further in
the forward direction; Trying to follow through on this behaviour, having in mind
the operational wisdom concerning real-world missions and experiences,
it is much safer to give priority to the component that manoeuvres the AUV to the side,
in order for the AUV to be able to circumnavigate the obstacles it might be heading towards.
f⊥ = $sign( f ⊥ ) min{| f ⊥ | , u/T, u (k − 1)/T + u̇}
f = min{u/T, u (k − 1)/T + u̇}2 − f ⊥ 2 − u ( k − 1) /T

— therefore, require the f ⊥ component to be as large as admissible,

and adjust the f accordingly
else
— otherwise the required dominant behaviour produced
by the virtual potentials framework is to decelerate or break the AUV;
Trying to follow through on this behaviour, it is much safer
to give priority to the component that decreases the forward speed of
the AUV, in order for the AUV to avoid colliding with possible obstacles
it is heading towards.
f = max{ f , − u (k − 1)/T, − u̇}
$
f ⊥ = sign( f ⊥ ) min{| f ⊥ | , (u̇/T )2 − ( f + u (k − 1)/T )2 }
— therefore, require that f component to be as large as
admissible, and adjust the f ⊥ accordingly
end if
end if
def
ψcrit = minimum right-hand side of (87) and (88)
if either of the left-hand sides of (87) or (88) > ψcrit
— if, after preceding adjustments, ∠ (f − f · u (k)/ u (k) ) is inadmissible
if f > 0
f = f cot ψcrit + f ⊥ cos2 ψcrit − (u (k − 1)/T ) sin2 ψcrit
— again, pay attention to the intention of the manoeuvre, and if the dominant behavior is
acceleration, decrease the accelerating component of f , f , further
end if
f ⊥ = sign( f ⊥ ) · ( f + u (k − 1)/T ) tan ψcrit
— clamp the f ⊥ component to an admissible value
end if

Table 3. Controlling Force Clamping / Saturation

6.2 Simulation 2
The second simulation presents a cruise in formation down a heavily cluttered corridor
defined by two larger rectangular obstacles. Figure 12 presents the actual paths traveled.
Figure 13 presents the speeds of the four vehicles.
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 127
29

550

500

450

400

350

x [m] 300

250

200

150

100

50

0
0 50 100 150 200 250 300 350 400 450 500 550
y [m]
Fig. 12. Paths of the 4 HILS models of AUVs based on the precursor to the NPS Aries vehicle
cruising in a cluttered environment.

u(t) [kts]

0 250 500 750 1000 1250

u(t) [kts]

0 250 500 750 1000 1250

u(t) [kts]

0 250 500 750 1000 1250

u(t) [kts]

0 250 500 750 1000 1250

Fig. 13. Speeds of the 4 HILS models of AUVs based on the precursor to the NPS Aries
vehicle cruising in a cluttered environment.
128
30 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

6.2.1 Cruise phase 1

A cruise is started in a formation. However, very soon the formation encounters the first
obstacle. As the leading formation-members are momentarily slowed down before they
circumnavigate to either side of the obstacle, the trailing members “pile up” in front of this
artificial potential barrier (especially the AUV closest to the origin). This is evident in the dips
and temporary confusion before the first, circular obstacle. However, the operational safety
approach that is implicitly encapsulated by the cross-layer design is preserved. The vehicles
break formation, so that one of the vehicle circumnavigates the first obstacle on the left, and
the others on the right. Since this produces a significantly different trajectory from the rest of
the group, vehicle 1 isn’t able to rejoin the formation until much later.

6.2.2 Cruise phase 2

The other 3 vehicles (2, 3 and 4), before being able to restore a formation encounter the first of
the two large rectangular obstacles. Note that vehicles 2 and 3 remain in the leader-follower
arrangement as evidenced by their closely matching trajectories in this phase of the cruise.
The “outrigger” vehicle 4, trying to keep in formation with 2 and 3, encounters the large
rectangular obstacle at a bearing much closer to head-on. Therefore, it executes a significant
course change manoeuvre, during which it cannot satisfactorily compromise between safe
avoidance of the obstacle and staying in formation with 2 and 3. As vehicles 2 and 3
navigate in formation through the strait in between the circular and the first large rectangular
obstacle, the leader vehicle 3 starts to manoeuvre to starboard towards the way-point. This
manoeuvre causes the formation cell vertex trailing behind vehicle 3 that represents the
dominant navigation goal for vehicle 2 to start accumulating speed in excess of what 2 is
able to match. This is due to the fact that as 3 swings to starboard, the formation cell vertex
“sweeps” through water with a velocity that consists of the sum of linear velocity of vehicle
3 and the tangential velocity contributed by the “arm” of the formation cell f . Therefore, the
formation is temporarily completely broken.

6.2.3 Cruise phase 3

However, the breaking of formation between 2 and 3 occurs at such a time that 1 catches up
with 2 before 2 gets much farther afield, presenting its trailing cell vertex as a local navigation
goal to 2. That is why 2 exhibits a hard break to starboard, trying to form itself up as a follower
of vehicle 1. However, just as 2 is completing its formation, vehicle 1, manoeuvres around the
final obstacle – the small diagonally presented rectangle. As the trailing cell vertex of 1 is, from
2’s viewpoint, shadowed by the obstacle’s repulsion, it reorients towards what until then is a
secondary navigation goal in its vicinity – the cell vertex of the “latecomer” of Phase 2, vehicle
4. This reorientation is what contributes to 2’s “decision” to circumnavigate the diagonal
rectangle to starboard, rather than to port, as would be optimal if no formation influences
were present. Phase 3 finishes as vehicle 2 is trying to pursue vehicle 4, and vehicle 4 corners
the diagonal rectangle, getting away from vehicle 2.

6.2.4 Cruise phase 4

Phase 4 is entered into without formations. This phase is characterized by converging on
the way-point, which all the vehicles reach independently, followed by re-establishing the
formation. However, an ideal formation is impossible due to operational safety, as no vehicle
is “willing” to approach the second large rectangle (towards the top of the figure). This is
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 129
31

exacerbated by reduced manoeuvring capabilities, as all vehicles reduce speed in the vicinity
of the way-point.

7. Conclusion
The chapter has presented a virtual potentials-based decentralized formation guidance
framework that operates in 2D. The framework guarantees the stability of trajectories,
convergence to the way-point which is the global navigation goal, and avoidance of salient,
hazardous obstacles. Additionally, the framework offers a cross-layer approach to subsuming
two competing behaviours that AUVs in a formation guidance framework need to combine
– a priority of formation maintenance, opposed by operational safety in avoiding obstacles
while cruising amidst clutter.
Additionally to the theoretical contribution, a well-rounded functional hardware-in-the-loop
system (HILS) for realistic simulative analysis was presented. Multiple layers of realistic
dynamic behaviour are featured in the system:
1. A full-state coupled model dynamics of a seaworthy, long-autonomy AUV model based
on rigid-body physics and hydrodynamics of viscous fluids like water,
2. An unbiased rate-limited white noise model of the process noise,
3. A non-stationary generator of measurement noise based on Gaussian Markov models with
an explicitly included fault-mode,
4. An outlier-elimination scheme based on the evaluation of the state estimate covariance
returned by the employed estimator,
5. A Scaled Unscented Transform Sigma-Point Kalman Filter (SP-UKF) that can work either
in the filtering mode, or a combination of filtering and pure-prediction mode when faulty
measurements are present, utilizing a full-state non-linear coupled AUV model dynamics,
6. A command signal adaptation mechanism that accents operational safety concerns by
prioritizing turning manoeuvres while accelerating, and “pure” braking / shedding
forward speed when decelerating.

7.1 Further work

Several distinct areas of research, based on the developed HILS framework, remain to
ascertain the quality of the presented virtual potential-based decentralized cooperative
framework. These are necessary in order to clear the framework for application in costly
and logistically demanding operations in the real Ocean environment.
1. Realistically model the representation of knowledge of the other AUVs aboard each AUV
locally.
This can be approached on several fronts:
(a) Exploring the realistic statistics of the sensing process when applied to sensing other
AUVs as opposed to salient obstacles in the waterspace. Exploring and modeling the
beam-forming issues arising with mechanically scanning sonars vs. more complex and
costlier multi-beam imaging sonars,

(b) Exploring the increases in complexity (and computer resource management),

numerical robustness and stability issues of AUV-local estimation of other AUVs in
the formation,
130
32 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

(c) Dealing with the issues of the instability of the “foreign” AUVs’ state estimates
covariance matrix by one of three ways: (i) using synchronous, pre-scheduled
hydroacoustic communication. Communication would entail improved estimates
coming from on-board the AUVs, where the estimates are corrected by collocated
measurement; (ii) exploring an on-demand handshake-based communication scheme.
Handshaking would be initiated by an AUV polling a team-member for a correction to
the local estimate featuring unacceptably large covariance; (iii) exploring a predictive
communication scheme where the AUVs themselves determine to broadcast their
measurements without being polled. This last option needs to involve each AUV
continually predicting how well other AUVs are keeping track of its own state
estimates.
2. Explore the applicability of the framework to non-conservative, energetic manoeuvring
in 3D, i.e. use the same framework to generate commands for the depth / pitch
low-level controllers. Explore the behaviour of 3D-formations based on the honeycombs
(3D tesselations) of the vector space of reals.

8. References
Abkowitz, M. (1969). Stability and Motion Control of Ocean Vehicles, MIT Press Cambridge MA
USA.
Ahn, S., Rauh, W. & Recknagel, M. (1999). Ellipse fitting and parameter assessment of
circular object targets for robot vision, Intelligent Robots and Systems 1999. IROS ’99
Proceedings. 1999 IEEE/RSJ International Conference on, Vol. 1, pp. 525 –530.
Allen, B., Stokey, R., Austin, T., Forrester, N., Goldborough, R., Purcell, M. & von Alt, C. (1997).
REMUS A Small Low Cost AUV system description field trials and performance
results, OCEANS 1997 Conference Record (IEEE), Vol. 2, pp. 994–100.
Allen, T., Buss, A. & Sanchez, S. (2004). Assessing obstacle location accuracy in the REMUS
unmanned underwater vehicle, Proceedings - Winter Simulation Conference, Vol. 1,
pp. 940–948.
An, P., Healey, A., Park, J. & Smith, S. (1997). Asynchronous data fusion for auv navigation via
heuristic fuzzy filtering techniques, OCEANS ’97. MTS/IEEE Conference Proceedings,
Vol. 1, pp. 397 –402 vol.1.
Barisic, M., Vukic, Z. & Miskovic, N. (2007a). Kinematic simulative analysis of virtual potential
field method for AUV trajectory planning, in K. Valavanis & Z. Kovacic (eds), 15th
Mediterranean Conference on Control and Automation, 2007. Proceedings of the, p. on CD.
Barisic, M., Vukic, Z. & Miskovic, N. (2007b). A kinematic virtual potentials trajectory planner
for AUV-s, in M. Devy (ed.), 6th IFAC Symposium on Intelligent Autonomous Vehicles,
2007. Proceedings of the, p. on CD.
Bildberg, R. (2009). Editor’s foreword, in B. R. (ed.), Proceedings of the 16th International
Symposium on Unmanned Untethered Submersibles Technology, Autonomous Undersea
Systems Institute. on CD.
Boncal, R. (1987). A Study of Model Based Maneuvering Controls for Autonomous Underwater
Vehicles, Master’s thesis, Naval Postgraduate School Monterey CA USA.
Carder, K., Costello, D., Warrior, H., Langebrake, L., Hou, W., Patten, J. & Kaltenbacher, E.
(2001). Ocean-science mission needs: Real-time AUV data for command control and
model inputs, IEEE Journal of Oceanic Engineering 26(4): 742–751.
Formation Guidance
Formation Guidance of AUVs Usingof AUVs Control
Decentralized UsingFunctions
Decentralized Control Functions 131
33

Clegg, D. & Peterson, M. (2003). User operational evaluation system of unmanned underwater
vehicles for very shallow water mine countermeasures, OCEANS 2003 Conference
Record (IEEE), Vol. 3, pp. 1417–1423.
Curtin, T., Bellingham, J., Catipovic, J. & Webb, D. (1993). Autonomous oceanographic
sampling networks, Oceanography 6(3): 86–94.
Duda, R. O. & Hart, P. E. (1972). Use of the Hough transformation to detect lines and curves
in pictures, Communications of the ACM 15(1): 11–15.
Eisman, D. (2003). Navy ships seize boats carrying mines in Iraqi port, The Virginian-pilot .
Farrell, J., S., P. & Li, W. (2005). Chemical plume tracing via an autonomous underwater
vehicle, IEEE Journal of Oceanic Engineering 30(2): 428–442.
Faucette, W. M. (1996). A geometric interpretation of the solution of the general quartic
polynomial, The American Mathematical Monthly 103(1): 51–57.
Gertler, M. & Hagen, G. (1967). Standard Equations of Motion for Submarine Simulation,
Technical report, David W. Taylor Naval Ship Research and Development Center
Bethesda MD USA.
Haule, D. & Malowany, A. (1989). Object recognition using fast adaptive Hough transform,
Communications Computers and Signal Processing 1989. Conference Proceedings IEEE
Pacific Rim Conference on, pp. 91 –94.
He, Y. & Li, Z. (2008). An effective approach for multi-rectangle detection, Young Computer
Scientists 2008. ICYCS. The 9th International Conference for, pp. 862 –867.
Healey, A. & Lienard, D. (1993). Multivariable sliding mode control for autonomous diving
and steering of unmanned underwater vehicles, Oceanic Engineering IEEE Journal of
18(3): 327 –339.
Hough, P. & Powell, B. (1960). A method for faster analysis of bubble chamber photographs,
Vol. 18, Italian Physical Society, pp. 1184 –1191.
Illingworth, J. & Kittler, J. (1987). The adaptive Hough transform, Pattern Analysis and Machine
Intelligence IEEE Transactions on PAMI-9(5): 690 –698.
Jiang, X., Huang, X., Jie, M. & Yin, H. (2007). Rock detection based on 2D maximum entropy
thresholding segmentation and ellipse fitting, Robotics and Bioimetics 2007. ROBIO
2007. IEEE International Conference on, pp. 1143 –1147.
Jung, C. & Schramm, R. (2004). Rectangle detection based on a windowed Hough transform,
Computer Graphics and Image Processing 2004. Proceedings, 17th Brazilian Symposium on,
pp. 113 – 120.
Lienard, D. (1990). Sliding Mode Control for Multivariable AUV Autopilots, Master’s thesis, Naval
Postgraduate School Monterey CA USA.
Maitre, H. (1986). Contribution to the prediction of performances of the Hough transform,
Pattern Analysis and Machine Intelligence IEEE Transactions on PAMI-8(5): 669 –674.
Marco, D. & Healey, A. (2000). Current developments in underwater vehicle control
and navigation: the NPS Aries AUV, OCEANS Conference record (IEEE), Vol. 2,
pp. 1011–1016.
Marco, D. & Healey, A. (2001). Command control and navigation experimental results with
the NPS Aries AUV, IEEE Journal of Oceanic Engineering 26(4): 466–476.
Nguyen, T., Pham, X. & Jeon, J. (2009). Rectangular object tracking based on standard Hough
transform, Robotics and Biomimetics 2008. ROBIO IEEE International Conference on,
pp. 2098 –2103.
Pang, S. (2006). Development of a guidance system for AUV chemical plume tracing, OCEANS
2006 Conference Record (IEEE), pp. 1–6.
132
34 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

Pang, S., Arrieta, R., Farell, J. & Li, W. (2003). AUV reactive planning: Deepest Point, OCEANS
2009 Conference Record (IEEE), Vol. 4, pp. 2222–2226.
Papoulias, F., Cristi, R., Marco, D. & Healey, A. (1989). Modeling, sliding mode control design,
and visual simulation of auv dive plane dynamic response, Unmanned Untethered
Submersible Technology, 1989. Proceedings of the 6th International Symposium on, pp. 536
–547.
Pilu, M., Fitzgibbon, A. & Fisher, R. (1996). Ellipse-specific direct least-square fitting, Image
Processing 1996. Proceedings. International Conference on, Vol. 3, pp. 599 –602.
Plueddemann, A., Packard, G., Lord, J. & Whelan, S. (2008). Observing Arctic Coastal
Hydrography Using the REMUS AUV, Autonomous Underwater Vehicles 2008. AUV
2008. IEEE/OES, pp. 1–4.
Rizon, M., Yazid, H. & Saad, P. (2007). A comparison of circular object detection using Hough
transform and chord intersection, Geometric Modeling and Imaging 2007. GMAI ’07,
pp. 115 –120.
Stewart, I. (2003). Galois Theory, Chapman & Hall / CRC Press, 6000 Broken Sound Pkwy.
Suite 300 Boca Raton FL 33487 USA.
Tuohy, S. T. (1994). A simulation model for AUV navigation, Proceedings of the 1994 Symposium
on Autonomous Underwater Vehicle Technology, pp. 470–478.
US Navy (2004). The US Navy Unmanned Underwater Vehicles Master Plan,
http://www.navy.mil/navydata/technology/uuvmp.pdf.
van der Merwe, R. (2004). Sigma-Point Kalman Filters for Probabilistic Inference in Dynamic
State-Space Models, PhD thesis, OGI School of Science & Engineering, Oregon Health
& Science University, Portland, OR, USA.
 6

Modeling and Motion Control

Strategy for AUV
Lei Wan and Fang Wang
Harbin Engineering University
China

1. Introduction
Autonomous Underwater Vehicles (AUV) speed and position control systems are subjected
to an increased focus with respect to performance and safety due to their increased number
of commercial and military application as well as research challenges in past decades,
including underwater resources exploration, oceanographic mapping, undersea wreckage
salvage, cable laying, geographical survey, coastal and offshore structure inspection, harbor
security inspection, mining and mining countermeasures (Fossen, 2002). It is obvious that all
kinds of ocean activities will be greatly enhanced by the development of an intelligent
underwater work system, which imposes stricter requirements on the control system of
underwater vehicles. The control needs to be intelligent enough to gather information from
the environment and to develop its own control strategies without human intervention
(Yuh, 1990; Venugopal and Sudhakar, 1992).
However, underwater vehicle dynamics is strongly coupled and highly nonlinear due to
added hydrodynamic mass, lift and drag forces acting on the vehicle. And engineering
problems associated with the high density, non-uniform and unstructured seawater
environment, and the nonlinear response of vehicles make a high degree of autonomy
difficult to achieve. Hence six degree of freedom vehicle modeling and simulation are quite
important and useful in the development of undersea vehicle control systems (Yuh, 1990;
Fossen 1991, Li et al., 2005). Used in a highly hazardous and unknown environment, the
autonomy of AUV is the key to work assignments. As one of the most important subsystems
of underwater vehicles, motion control architecture is a framework that manages both the
sensorial and actuator systems (Gan et al., 2006), thus enabling the robot to undertake a
user-specified mission.
In this chapter, a general form of mathematical model for describing the nonlinear vehicle
systems is derived, which is powerful enough to be applied to a large number of
underwater vehicles according to the physical properties of vehicle itself to simplify the
model. Based on this model, a simulation platform “AUV-XX” is established to test motion
characteristics of the vehicle. The motion control system including position, speed and
depth control was investigated for different task assignments of vehicles. An improved S-
surface control based on capacitor model was developed, which can provide flexible gain
selections with clear physical meaning. Results of motion control on simulation platform
“AUV-XX” are described.
134 Autonomous Underwater Vehicles

2. Mathematical modeling and simulation

Six degree of freedom vehicle simulations are quite important and useful in the
development of undersea vehicle control systems. There are several processes to be modeled
in the simulation including the vehicle hydrodynamics, rigid body dynamics, and actuator
dynamics, etc.

2.1 AUV kinematics and dynamics

The mathematical models of marine vehicles consist of kinematic and dynamic part, where
the kinematic model gives the relationship between speeds in a body-fixed frame and
derivatives of positions and angles in an Earth-fixed frame, see Fig.1. The vector of positions
T
and angles of an underwater vehicle η = [ x , y , z , ϕ , θ ,ψ ] is defined in the Earth-fixed
coordinate system(E) and vector of linear and angular v v = [u , v , w , p , q , r ]T elocities is
defined in a body-fixed(B) coordinate system, representing surge, sway, heave, roll, pitch
and yaw velocity, respectively.

Fig. 1. Earth-fixed and body-fixed reference frames

According to the Newton-Euler formulation, the 6 DOF rigid-body equations of motion in
the body-fixed coordinate frame can be expressed as:

r r (G )
⎧m ⎡(u − v r + w q ) − x q 2 + r 2 + y ( pq − r ) + z ( pr + q )⎤ = X
⎪ ⎣ r G G ⎦
⎪ ⎡
⎪m ⎣( v r − wr p + ur r ) − yG (r + p ) + zG (qr − p ) + xG (qp + r )⎤⎦ = Y
2 2

⎪
⎪m ⎡⎣( w  r − ur q + vr p) − zG ( p2 + q 2 ) + xG (rp − q ) + yG (rq + p )⎤⎦ = Z
⎨ (1)
⎪ I x p + ( I z − I y )qr + m ⎡ yG ( w  r + pvr − qur ) − zG ( v r + rur − pwr )⎤⎦ = K
⎪ ⎣
⎪ I y q + ( I x − I z )rp + m ⎡ zG ( u r + wr q − vr r ) − xG ( w
 r + pvr − ur q ) ⎤⎦ = M
⎪ ⎣
⎪ I zr + ( I y − I z )pq + m ⎡ xG ( v r + ur r − pwr ) − yG ( u r + qwr − vr r ) ⎤ = N
⎩ ⎣ ⎦

where m is the mass of the vehicle, I x , I y and I z are the moments of inertia about the
xb , yb and zb -axes, x g , y g and z g are the location of center of gravity, ur ,vr ,wr are relative
Modeling and Motion Control Strategy for AUV 135

translational velocities associated with surge, sway and heave to ocean current in the body-
fixed frame, here assuming the sea current to be constant with orientation in yaw only,
which can be described by the vector U c = [uc , vc , wc ,0,0,α c ]T . The resultant
forces X , Y , Z , K , M , N includes positive buoyant B − W = ΔP ( since it is convenient to
design underwater vehicles with positive buoyant such that the vehicle will surface
automatically in the case of an emergency), hydrodynamic forces X H ,YH , ZH , K H , M H , N H
and thruster forces.

2.2 Thrust hydrodynamics modeling

The modeling of thruster is usually done in terms of advance ratio J 0 , thrust
coefficients KT and torque coefficient KQ . By carrying out an open water test or a towing
tank test, a unique curve where J 0 is plotted against KT and KQ can be obtained for each
propeller to depict its performance. And the relationship of the measured thrust force versus
propeller revolutions for different speeds of advance is usually least-squares fitting to a
quadratic model.
Here we introduce a second experimental method to modeling thruster dynamics. Fig.2
shows experimental results of thrusters from an open water test in the towing tank of the
Key Lab of Autonomous Underwater Vehicles in Harbin Engineering University. The
results are not presented in the conventional way with the thrust coefficient KT plotted
versus the open water advance coefficient J 0 , for which the measured thrust is plotted as a
function of different speeds of vehicle and voltages of the propellers.
The thrust force of the specified speed of vehicle under a certain voltage can be finally
approximated by Atiken interpolation twice. In the first interpolation, for a certain voltage,
the thrust forces with different speeds of the vehicle (e.g. 0m/s, 0.5m/s, 1.0m/s, 1.5m/s) can
be interpolated from Fig.1, and plot it versus different speeds under a certain voltage. Then
based on the results of the first interpolation, for the second Atiken interpolation we can
find the thrust force for the specified speed of the vehicle.

Fig. 2. Measured thrust force as a function of propeller driving voltage for different speeds
of vehicle
136 Autonomous Underwater Vehicles

Compared with conventional procedure to obtain thrust that is usually done firstly by linear
approximating or least-squares fitting to KT - J 0 plot (open water results), then using
formulation Ft = Kt n 2 D 4 to compute the thrust Ft . The experimental results of open water
can be directly used to calculate thrust force without using the formulation, which also can
be applied to control surface of rudders or wings, etc.

2.3 General dynamic model

To provide a form that will be suitable for simulation and control purposes, some
rearrangements of terms in Eq.(1) are required. First, all the non-inertial terms which have
velocity components were combined with the fluid motion forces and moments into a fluid
vector denoted by the subscript vis (viscous). Next, the mass matrix consisting all the
coefficient of rigid body’s inertial and added inertial terms with vehicle acceleration
components u , v , w
 , p , q , r was defined by matrix E , and all the remaining terms were
combined into a vector denoted by the subscript else, to produce the final form of the model:

EX = Fvis + Felse + Ft (2)

where X = [u , v , w , p , q , r ]T is the velocity vector of vehicle with respect to the body-fixed

frame.
Hence, the 6 DOF equations of motion for underwater vehicles yield the following general
representation:

⎧⎪ X = E −1 ( Fvis + Felse + Ft )
⎨ (3)
⎪⎩η = J (η ) X
with

⎡m− Xu 0 0 0 mzG −myG ⎤

⎢ ⎥
⎢ 0 m −Y v 0 − mz G −Yp 0 mx G −Yr ⎥
⎢ 0 0 m−Zw myG −mxG −Zq 0 ⎥
E=⎢ ⎥ (4)
⎢ 0 −mzG −K v 0 I x −K p 0 −Kr ⎥
⎢ ⎥
⎢ mzG 0 −mxG − M w 0 I y − Mq 0 ⎥
⎢ ⎥
⎣⎢ 0 mx G − N 
v 0 − N p 0 I z − N r ⎦
 ⎥
where J (η ) is the transform matrix from body-fixed frame to earth-fixed frame, η is the
vector of positions and attitudes of the vehicle in earth-fixed frame.
The general dynamic model is powerful enough to apply it to different kinds of underwater
vehicles according to its own physical properties, such as planes of symmetry of body,
available degrees of freedom to control, and actuator configuration, which can provide an
effective test tool for the control design of vehicles.

3. Motion control strategy

In this section, the design of motion control system of AUV-XX is described. The control
system can be cast as two separate designs, which include both position and speed control
Modeling and Motion Control Strategy for AUV 137

in horizontal plane and the combined heave and pitch control for dive in vertical plane. And
an improved S-surface control algorithm based on capacitor plate model is developed.

3.1 Control algorithm

As a nonlinear function method to construct the controller, S-surface control has been
proven quite effective in sea trial for motion control of AUV in Harbin Engineering
University (Li et al., 2002). The nonlinear function of S-surface is given as:

u = 2.0 ( 1.0 + exp( − k1 e − k2 e )) − 1.0 (5)

where e , e are control inputs, and they represent the normalized error and change rate of
the error, respectively; u is the normalized output in each degree of freedom; k1 , k2 are
control parameters corresponding to control inputs e and e respectively, and we only need
to adjust them to meet different control requirements.
Based on the experiences of sea trials, the control parameters k1 , k2 can be manually
adjusted to meet the fundamental control requirements, however, whichever combination of
k1 , k2 we can adjust, it merely functions a global tuning which dose not change control
structure. Here the improved S-surface control algorithm is developed based on the
capacitor with each couple of plates putting restrictions on the control variables e , e
respectively, which can provide flexible gain selection with proper physical meaning.

Fig. 3. Capacitor plate model

The capacitor plate model as shown in Fig.3 demonstrates the motion of a charged particle
driven by electrical field in capacitor is coincident with the motion of a controlled vehicle
from current point ( e , e ) to the desired point, for which the capacitor plate with voltage
138 Autonomous Underwater Vehicles

serves as the controller, and the equilibrium point of electrical field is the desired position
that the vehicle is supposed to reach.
Due to the restriction of two couples of capacitor plates put on control variables e and e , the
output of model can be obtained as

+U0 −U0
y=u +u = F(L1 , L2 )( +U0 ) + F(L2 , L1 )( −U0 ) (6)

where L1 , L2 are horizontal distances from the current position of the vehicle to each
capacitor plate, respectively, and the restriction function F(*,*) is defined to be hyperbolic
function of L1 , L2 by Ren and Li (2005):

⎧ L−1k
⎪F(L1 , L2 ) = − k
⎪ L1 + L−2k
⎨ (7)
−k
⎪F(L , L ) = L2
⎪ 2 1
L1 + L−2k
−k
⎩
The restriction function F(*,*) reflects the closer the current position ( e , e ) of vehicle moving
to capacitor plate, the stronger the electrical field is. Choosing U0 = 1 , the output of
capacitor plate model yields:

L−1k − L−2k ( e + e)k − ( e0 − e )k

u= −k −k
U0 = 0 (8)
L1 + L2 ( e0 + e )k + ( e0 − e )k

where e0 is the distance between the plate and field equilibrium point of capacitor.
An improved S-surface controller based on the capacitor plate model is proposed, that is

⎧ ⎛ e0 − ei ki 1 ⎞
⎪uei = [2.0 ⎜ 1.0 + ( ) ⎟ − 1.0]
⎪ ⎝ e0 + ei ⎠
⎪
⎪ ⎛ e − 
e i ki 2 ⎞
⎨uei = [2.0 ⎜ 1.0 + ( ) ⎟ − 1.0]
0
(9)
⎪ ⎝ e0 + ei
 ⎠
⎪ f = K ⋅u + K ⋅u
 
⎪ i ei ei ei ei
⎪⎩

where f i is the outputted thrust force of controller for each DOF, and K ei = K ei = K i is the
maximal thrust force in i th DOF, therefore the control output can be reduced to

⎧ ⎛ e0 − ei ki 1 ⎞ ⎛ e − e ⎞
⎪ui = uei + uei = [2.0 ⎜ 1.0 + ( ) ⎟ − 1.0] + [2.0 ⎜ 1.0 + ( 0 i )ki 2 ⎟ − 1.0]
⎨ ⎝ e0 + ei ⎠ ⎝ e0 + ei ⎠ (10)
⎪ f = K ×u
⎩ i i i

The capacitor model’s S-surface control can provide flexible gain selections with different
forms of restriction function to L1 , L2 to meet different control requirements for different
phases of control procedure.
Modeling and Motion Control Strategy for AUV 139

3.2 Speed and position control in horizontal plane

Since AUV-XX is equipped with two transverse tunnel thrusters in the vehicle fore and aft
respectively and two main thrusters (starboard and port) aft in horizontal plane, which can
produce a force in the x-direction needed for transit and a force in the y-direction for
maneuvering, respectively. So both speed and position controllers are designed in
horizontal plane.
Speed control is to track the desired surge velocity with fixed yaw angle and depth, which is
usually used in long distance transfer of underwater vehicles. Before completing certain
kind of undersea tasks, the vehicle needs to experience long traveling to achieve the
destination. In this chapter, speed control is referred to a forward speed controller in surge
based on the control algorithm we introduced in above section, its objective is to make the
vehicle transmit at a desired velocity with good and stable attitudes such as fixed yaw and
depth.

Fig. 4. Position and speed control loop

Position control enables the vehicle to perform various position-keeping functions, such as
maintaining a steady position to perform a particular task, following a prescribed trajectory
to search for missing or seek after objects. Accurate position control is highly desirable when
the vehicle is performing underwater tasks such as cable laying, dam security inspection
and mine clearing. To ensure AUV-XX to complete work assignments of obstacles avoiding,
target recognition, and mine countermeasures, we design position controllers for surge,
sway, yaw and depth respectively for equipping the vehicle with abilities of diving at fixed
deepness, navigating at desired direction, sailing to given points and following the given
track, etc.
As for the desired or target position or speed in the control system, it is the path planning
system who decides when to adopt and switch control scheme between position and speed,
the desired position that the vehicle is supposed to reach, and the velocity at which the
140 Autonomous Underwater Vehicles

vehicle should navigate with respect to the present tasks and motion states of the vehicle as
well as operation environment.
Fig.4 shows both position and speed control procedures. It can be seen that it is easier to
realize the control algorithm. For position control of i th DOF, the control inputs are the
position error and the change rate of position error, that is the velocity obtained from
motion sensors; while for speed control, the velocity error and acceleration are control
inputs, since AUV-XX is not equipped with IGS to acquire the acceleration of the vehicle,
acceleration is calculated differential the velocity in each control step.

3.3 Combined control of pitch and heave in vertical plane

Since when the vehicle is moving at a high speed, the thrust that tunnel thrusters can
provide will strongly degrade, it is difficult to control the depth merely using tunnel
thrusters. Considering that once the depth or height of the vehicle changes, the pitch will
change with it, and vice versa, so we combine pitch with heave control for diving when the
vehicle is moving at some high speed to compensate for the thrust reduction of tunnel
thrusters. In that case, the desired pitch angle can be designed as a function of the surge
velocity of the vehicle:

⎧ kθ ⋅ Δz
− − θ0 u ≥ 0.8m / s
θT = ⎪⎨ u (11)
⎪0 u < 0.8m / s
⎩

where θT is the target pitch angle, kθ is a positive parameter to be adjusted, Δz is the

depth deviation, u is the surge velocity of vehicle, θ0 is the pitch angle when the vehicle is
in static equilibrium. Since the change of pitch angle is usually associated with depth change
and they affect each other, the target pitch is the output of the proportional control with
respect to depth change with the proportional parameters kθ , the velocity threshold of
0.8m/s is chosen based on the engineering experience of the sea trial and the capability of
the thruster system of the vehicle.
When the vehicle is moving at a low speed( u < 0.8m / s ), tunnel thrusters can normally
provide the needed force, so the command of target pitch will not be sent to motion control.
According to control law (9), we can get the output of the heave controller. Since the vehicle
is usually designed with positive buoyancy, the final output of control of heave can be
obtained by

f 3 = K 3 ⋅ u3 + ΔP (12)

where ΔP is the positive buoyancy.

And as the surge velocity grows, the thrust of vertical tunnel will experience worse
reduction and degradation, so the role it plays in the depth control will be greatly abridged,
as a result, the output of pitch control of the main vertical thrusters aft of the vehicle should
compensate for that, so the output of control will finally yields as

⎧⎪K 5u5 u ≤ 0.8

f5 = ⎨ (13)
⎩⎪K 5 (ε 1 ⋅ u3 + ε 2 ⋅ u5 ) u > 0.8
Modeling and Motion Control Strategy for AUV 141

with

⎧⎪ε 1 = α 1 / β u
⎨ (14)
⎪⎩ε 2 = α 2 ⋅ exp(u ⋅ u /10)

where α 1 , α 2 , β can be manually adjusted based on experiences. The block diagram of

combined control of pitch and depth in vertical plane can be in Fig.5.

Fig. 5. Combined control of pitch and depth in vertical plane

Fig. 6. AUV-XX simulation platform

4. Simulation results and analysis

To verify the feasibility and effectiveness of the motion control system for the vehicle AUV-
XX, simulations are carried out in the AUV-XX simulation platform. The vehicle researched
in this chapter named by AUV-XX, AUV-XX's configuration is basically a long cylinder of
0.5m in diameter and 5m in length with crossed type wings near its rear end. On each edge
of the wings, a thruster is mounted, which is used for both turn and dive. AUV-XX is also
equipped with a couple of lateral tunnel thrusters for sway and a couple of vertical tunnel
thrusters fore and aft of the vehicle, respectively. Based on the modeling method described
in above section, we established the AUV-XX simulation platform to carry out fundamental
142 Autonomous Underwater Vehicles

tests on its motion characteristics, stability and controllability. The states of the vehicle
including positions, attitudes and velocities are obtained at each instant by solving the
mathematical model equation by integration using a time step of 0.5s. Fig.6 shows the
interface of AUV-XX simulation platform. Fig.7 shows the data flow of the simulation
platform connecting with motion control system.
Figs.8–10 show the simulation results of capacitor plate model’s S-surface control for
separate position and speed control in surge, sway and yaw respectively, and also combined
control of heave and pitch for diving of the vehicle. The roll is left uncontrolled. And all the
vehicle states, including speed, are initialized to zero at the beginning of the each of the
simulation. The solid lines denote the actual responses of the vehicle while the dashed
denote the desired position or speed that the vehicle is commanded to achieve.

Fig. 7. Data flow of AUV-XX motion control in simulation platform

It can be seen from Fig.8 that, the vehicle is commanded to move to some specified positions
in surge, sway and yaw, respectively. For the case of surge as shown in Fig.8(a) and Fig.8(b),
the larger position deviation from the target position produces faster response of surge
velocity. Compared with responses of surge and yaw, the sway is slower with a rise time of
150s for the desired position is 16m, which may result from that the lateral resistance is
much larger longitudinally and the thrust of transverse tunnel thrusters can provide is
smaller than the main aft thrusters.
Fig.9 shows the speed control results of surge with constant yaw and depth keeping. The
desired velocity is 1m/s and once the vehicle is moving stably with such speed then the
vehicle is commanded to track the specified depth (5m) and yaw (45o) commands. It can be
seen that there is no overshoot in surge speed and system response is fast with the yaw
experiencing an accepted overshoot of ± 2o, and the depth is stably maintained, hence the
vehicle is able to move at a desired speed with a desired fixed heading and depth. The speed
control simulation results prove the feasibility of the proposed speed control strategy.
Modeling and Motion Control Strategy for AUV 143

(a) Surge displacement (b) Surge velocity response

(c) Sway displacement (d) Yaw displacement

Fig. 8. Position control results of surge, sway and yaw

144 Autonomous Underwater Vehicles

(a) Surge speed response (b) Yaw keep response

(c) Depth keep response

Fig. 9. Speed control results of surge with yaw and depth keeping
Fig.10 shows the combined control of heave and pitch for diving in vertical plane. For this
case, the velocity of surge that the vehicle is commanded to track is 1.5m/s, which is not
very large so that the vertical tunnel thrusters will suffer thrust reduction to some extent but
still can work to provide a portion of vertical thrust for diving. Hence when the vehicle is
commanded to dive, the pitch will not experience a large change, which is reasonable design
consideration in the case of large inertial vehicles.
Modeling and Motion Control Strategy for AUV 145

(c) Depth keep response (d) Pitch response

Fig. 10. Combined depth control of heave and pitch

5. Conclusions
In this chapter, the design of motion control system for Autonomous Underwater Vehicles is
described, which includes both position and speed control in horizontal plane and
combined control of heave and pitch in vertical plane. To construct the control system, a 6
DOF general mathematical model of underwater vehicles was derived, which is powerful
enough to apply it to different kinds of underwater vehicles according to its own physical
properties. Based on the general mathematical model, a simulation platform was established
to test motion characteristics, stability and controllability of the vehicle. To demonstrate the
performance of the designed controller, simulations have been carried out on AUV-XX
simulation platform and the capacitor plate model S-surface control shows a good
performance.

6. References
Fossen T. I. (1991). Nonlinear modeling and control of underwater vehicles. Ph.D. thesis,
Norwegian Institute of Technology-NTH, Trondheim, Norway
Fossen T. I. (2002). Marine control system: guidance, navigation and control of ships, rigs and
underwater vehicles. Marine Cybernetics, Trondheim, 254-260
Gan Yong, Sun Yushan, Wan Lei, Pang Yongjie (2006). Motion control system architecture of
underwater robot. Proceedings of the 6th World Congress on Intelligent Control and
Automation, Dalian, China, 8876-8880
Gan Yong, Sun Yushan, Wan Lei, Pang Yongjie (2006). Motion control system of underwater
robot without rudder and wing. Proceeding of the 2006 IEEE International Conference
on Intelligent Robotics and Systems, Beijing, China, 3006-3011
Li Xuemin, Xu Yuru (2002). S-control of automatic underwater vehicles. The Ocean
Engineering, 19(3), 81-84
146 Autonomous Underwater Vehicles

Li Ye, Liu Jianchen, Shen Minxue (2005). Dynamics model of underwater robot motion
control in 6 degrees of freedom. Journey of Harbin Institute of Technology, 12(4), 456-
459
Ren Yongping, Li Shengyi (2005). Design method of a kind of new controller. Control and
Decision, 20(4), 471-474
Venugopal KP, Sudhakar R (1992). On-line learning control of autonomous underwater
vehicles using feedforward neural networks. IEEE Journal of Oceanic Engineering,
17(4), 308-319
Yuh J (1990). Modeling and control of underwater robotic vehicles. IEEE Transaction on
Systems, Man, and Cybernetics, 20(6), 1475-1483
 7

Fully Coupled 6 Degree-of-Freedom

Control of an Over-Actuated
Autonomous Underwater Vehicle
Matthew Kokegei, Fangpo He and Karl Sammut
Flinders University
Australia

1. Introduction
Unmanned underwater vehicles (UUVs) are increasingly being used by civilian and defence
operators for ever more complex and dangerous missions. This is due to the underlying
characteristics of safety and cost effectiveness when compared to manned vehicles. UUVs
require no human operator be subject to the conditions and dangers inherent in the
underwater environment that the vehicle is exposed to, and therefore the risk to human life
is greatly minimised or even removed. Cost effectiveness, in both time and financial
respects, comes from a much smaller vehicle not containing the various subsystems required
to sustain life whilst underwater, as well as smaller, less powerful actuators not placing the
same levels of stress and strain on the vehicles as compared to a manned vehicle. This leads
to a much smaller team required to undertake the regular maintenance needed to keep a
vehicle operational. Taking these two main factors into account, the progression from
manned vehicles to unmanned vehicles is a logical step within the oceanographic industry.
Within the broad class of UUVs are the remotely operated vehicles (ROVs) and the
autonomous underwater vehicles (AUVs). Both of these types of vehicles have been
successfully used in industry, and their fundamental differences determine which type of
vehicle is suited to a particular mission. The key difference between the two is that an ROV
requires a tether of some description back to a base station, whereas an AUV does not. This
tether connects the ROV to a human operator who can observe the current state of the
vehicle and therefore provide the control for the vehicle while it executes its mission. This
tether, depending on the configuration of the vehicle, can also provide the electricity to
power the vehicles actuators, sensors and various internal electronic systems.
AUVs have an advantage over ROVs of not requiring this tether, which leads to two main
benefits. Firstly, an AUV requires little or no human interaction while the vehicle is
executing its mission. The vehicle is pre-programmed with the desired mission objectives
and, upon launch, attempts to complete these objectives without intervention from
personnel located at the base station. This minimises the effect of human error while the
vehicle is operational. The second benefit is the increased manoeuvrability that is possible
without a cable continuously attached to the AUV. This tether has the potential to become
caught on underwater structures, which could limit the possible working environments of
an ROV, as well as cause drag on the motion of the vehicle, thus affecting its manoeuvring
148 Autonomous Underwater Vehicles

performance. The possible range of the vehicle from the base station is also restricted,
depending on the length of this tether. These two main benefits of AUVs over ROVs lead, in
principle, to autonomous vehicles being selected for survey tasks in complex, dynamic and
dangerous underwater environments, and therefore AUVs are the subject of this chapter.
Furthermore, combining the desired performance characteristics of AUVs, with the
aforementioned complex operation environment, leads to the conclusion that controllers
implemented within AUVs must be precise and accurate, as well as robust to disturbances
and uncertainties. Hence, the focus of this chapter will primarily be on the precise and
robust control of AUVs.
Within the autonomy architecture of AUVs are three main systems. These are:- the guidance
system, which is responsible for generating the trajectory for the vehicle to follow; the
navigation system, which produces an estimate of the current state of the vehicle; and the
control system, which calculates and applies the appropriate forces to manoeuvre the vehicle
(Fossen, 2002). This chapter will focus on the control system and its two principal
subsystems, namely the control law and the control allocation.
The chapter will be divided into three main parts with the first part focusing on the design
and analysis of the control law, the second looking at control allocation, and the third
providing an example of how these two systems combine to form the overall control system.
Within the control law design and analysis section, the requirements of how the various
systems within an AUV interact will be considered, paying particular attention to how these
systems relate to the control system. An overview of the underwater environment will be
given, which depicts the complexity of the possible disturbances acting on a vehicle. This
will be followed by an analysis of the equations of motion, namely the kinematic and kinetic
equations of motion that determine how a rigid body moves through a fluid. A summary of
the relevant frames of reference used within the setting of underwater vehicles will be
included. A review of the control laws that are typically used within the context of
underwater vehicles will be conducted to conclude this section of the chapter.
The second section will look at the role of control allocation in distributing the desired
control forces across a vehicle’s actuators. An analysis of the principal types of actuators
currently available to underwater vehicles will be conducted, outlining their useful
properties, as well as their limitations. This section will conclude with an overview of
various techniques for performing control allocation, with varying degrees of computational
complexity.
The third and final section of this chapter will present an example of an overall control
system for implementation within the architecture of AUVs. This example will demonstrate
how the control law and control allocation subsystems interact to obtain the desired
trajectory tracking performance while making use of the various actuators on the vehicle.

2. Control law design and analysis

Before delving into the laws governing how a particular control system produces a
correcting signal, it is necessary to look at the requirements of the various systems within an
AUV. This will provide an understanding of how the control system fits in with respect to
the overall autonomy architecture of an AUV. The different types of disturbances must be
acknowledged such that the effect of these disturbances is minimised, and the equations
related to the dynamic motion of the vehicle must be analysed. Only after reviewing these
factors can the control law be designed and analysed.
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 149

2.1 Requirements
As previously stated, the various components that make up the autonomy architecture of
AUVs are the guidance system, navigation system, and control system. All three of these
systems have their own individual tasks to complete, yet must also work cooperatively in
order to reliably allow a vehicle to complete its objectives. Figure 1 shows a block diagram
of how these various systems interact.

Fig. 1. Guidance, Navigation and Control Block Diagram.

2.1.1 Guidance
The guidance system is responsible for producing the desired trajectory for the vehicle to
follow. This task is completed by taking the desired waypoints defined pre-mission and,
with the possible inclusion of external environmental disturbances, generates a path for the
vehicle to follow in order to reach each successive waypoint (Fossen, 1994, 2002).
Information regarding the current condition of the vehicle, such as actuator configuration
and possible failures, can also be utilised to provide a realistic trajectory for the vehicle to
follow. This trajectory then forms the desired state of the vehicle, as it contains the desired
position, orientation, velocity, and acceleration information.

2.1.2 Navigation
The navigation system addresses the task of determining the current state of the vehicle. For
surface, land and airborne vehicles, global positioning system (GPS) is readily available and
is often used to provide continuous accurate positioning information to the navigation
system. However, due to the extremely limited propagation of these signals through water,
GPS is largely unavailable for underwater vehicles. The task of the navigation system is then
to compute a best estimate of the current state of the vehicle based on multiple
measurements from other proprioceptive and exteroceptive sensors, and to use GPS only
when it is available. This is completed by using some form of sensor fusion technique, such
as Kalman filtering or Particle filtering (Lammas et al., 2010, 2008), to obtain a best estimate
of the current operating condition, and allow for inclusion of a correction mechanism when
GPS is available, such as when the vehicle is surfaced. Overall, the task of the navigation
system is to provide a best estimate of the current state of the vehicle, regardless of what
sensor information is available.
150 Autonomous Underwater Vehicles

2.1.3 Control
The control system is responsible for providing the corrective signals to enable the vehicle to
follow a desired path. This is achieved by receiving the desired state of the vehicle from the
guidance system, and the current state of the vehicle from the navigation system. The
control system then calculates and applies a correcting force, through use of the various
actuators on the vehicle, to minimise the difference between desired and current states
(Fossen, 1994, 2002). This allows the vehicle to track a desired trajectory even in the presence
of unknown disturbances. Even though each of the aforementioned systems is responsible
for their own task, they must also work collaboratively to fully achieve autonomy within an
underwater vehicle setting.

2.2 Environment
Underwater environments can be extremely complex and highly dynamic, making the
control of an AUV a highly challenging task. Such disturbances as currents and waves are
ever present and must be acknowledged in order for an AUV to traverse such an
environment.

2.2.1 Currents
Ocean currents, the large scale movement of water, are caused by many sources. One
component of the current present in the upper layer of the ocean is due to atmospheric wind
conditions at the sea surface. Differing water densities, caused by combining the effect of
variation of salinity levels with the exchange of heat that occurs at the ocean surface, cause
additional currents known as thermohaline currents, to exist within the ocean. Coriolis
forces, forces due to the rotation of the Earth about its axis, also induce ocean currents, while
gravitational forces due to other planetary objects, such as the moon and the sun, produce
yet another effect on ocean currents (Fossen, 1994, 2002). Combining all of these sources of
water current, with the unique geographic topography that are present within isolated
coastal regions, leads to highly dynamic and complex currents existing within the world’s
oceans.

2.2.2 Wind generated waves

There are many factors that lead to the formation of wind generated waves in the ocean.
Wind speed, area over which the wind is blowing, duration of wind influencing the ocean
surface, and water depth, are just some of the elements that lead to the formation of waves.
Due to the oscillatory motion of these waves on the surface, any vehicle on the surface will
experience this same oscillatory disturbance. Moreover, an underwater vehicle will
experience both translational forces and rotational moments while at or near the surface due
to this wave motion.

2.3 Dynamics
All matter that exists in our universe must adhere to certain differential equations
determining its motion. By analysing the physical properties of an AUV, a set of equations
can be derived that determine the motion of this vehicle through a fluid, such as water. To
assist in reducing the complexities of these equations, certain frames of reference are utilised
depending on the properties that each frame of reference possesses. In order to make use of
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 151

these different reference frames for different purposes, the process of transforming
information from one frame to another must be conducted.

2.3.1 Frames of reference

Within the context of control systems, the two main reference frames used are the n-frame
and the b-frame. Both contain three translational components and three rotational
components, yet the origin of each frame differs. This difference in origin can lead to useful
properties which contain certain advantages when designing a control system.
2.3.1.1 N-Frame
The n-frame is a co-ordinate space usually defined as a plane oriented at a tangent to the
Earth’s surface. The most common of these frames for underwater vehicle control design is
the North-East-Down (NED) frame. As its name suggests, the three axes of the translational
components of this frame have the x-axis pointing towards true North, the y-axis pointing
towards East, and the z-axis in the downward direction perpendicular to the Earth’s surface.
In general, waypoints are defined with reference to a fixed point on the earth, and therefore
it is convenient to conduct guidance and navigation in this frame.
2.3.1.2 B-Frame
The b-frame, also known as the body frame, is a moving reference frame that has its origin
fixed to the body of a vehicle. Due to various properties that exist at different points within
the body of the vehicle, it is convenient to place the origin of this frame at one of these
points to take advantage of, for example, body symmetries, centre of gravity, or centre of
buoyancy. As a general rule, the x-axis of this frame points from aft to fore along the
longitudinal axis of the body, the y-axis points from port to starboard, and the z-axis points
from top to bottom. Due to the orientation of this frame, it is appropriate to express the
velocities of the vehicle in this frame.

2.3.2 Kinematic equation

As mentioned in 2.3.1, both the NED and body frames have properties that are useful for
underwater vehicle control design. Because both are used for different purposes, a means of
converting information from one frame to the other is required. The kinematic equation, (1),
achieves this task (Fossen, 1994, 2002).

η = J (η )υ
(1)
Here, the 6 degree-of-freedom (DoF) position and orientation vector in (1), decomposed in
the NED frame, is denoted by (2).

T
η = ⎡⎣ pn Θ ⎤⎦ (2)

Within (2), the three position components are given in (3),

p n = [ xn zn ]
T
yn (3)

and the three orientation components, also known as Euler angles, are given in (4).
152 Autonomous Underwater Vehicles

Θ = [φ θ ψ ]
T
(4)

The 6 DoF translational and rotational velocity vector in (1), decomposed in the body frame,
is denoted by (5).

T
ν = ⎡⎣ vob ωnb
b ⎤
⎦ (5)

Here, the three translational velocity components are given in (6),

vob = [ u v w]
T
(6)

and the three rotational velocity components are given in (7).

= [p q r]
b T
ωnb (7)

In order to rotate from one frame to the other, a transformation matrix is used in (1). This
transformation matrix is given in (8).

⎡ R n ( Θ ) 0 3× 3 ⎤
J (η ) = ⎢ b ⎥ (8)
⎣⎢ 0 3× 3 TΘ ( Θ ) ⎦⎥

Here, the transformation of the translational velocities from the body frame to the NED
frame are achieved by rotating the translational velocities in the body frame, (6), using the
Euler angles (4). Three principal rotation matrices are used in this operation, as shown in (9).

⎡1 0 0 ⎤ ⎡ cosθ 0 sin θ ⎤ ⎡ cosψ − sinψ 0⎤

⎢ ⎥ ⎢ ⎥ ⎢ ⎥
Rx ,φ = ⎢0 cos φ − sin φ ⎥ , Ry ,θ = ⎢ 0 1 0 ⎥ , Rz ,ψ = ⎢ sinψ cosψ 0⎥ (9)
⎣⎢0 sin φ cos φ ⎦⎥ ⎢⎣ − sin θ 0 cosθ ⎦⎥ ⎢⎣ 0 0 1 ⎦⎥

The order of rotation is not arbitrary, due to the compounding effect of the rotation order.
Within guidance and control, it is common to use the zyx-convention where rotation is
achieved using (10).

Rbn ( Θ ) = Rz ,ψ Ry ,θ Rx ,φ (10)

Overall, this yields the translational rotation matrix (11).

⎡ cosψ cosθ − sinψ cos φ + cosψ sin θ sin φ sinψ sin φ + cosψ sin θ cos φ ⎤
Rbn ( Θ ) = ⎢⎢ sinψ cosθ cosψ cos φ + sinψ sin θ sin φ
⎥
− cosψ sin φ + sinψ sin θ cos φ ⎥ (11)
⎢⎣ − sin θ cosθ sin φ cosθ cos φ ⎥⎦

The transformation of rotational velocities from the body frame to the NED frame is
achieved by again applying the principal rotation matrices of (9). For ease of understanding,
firstly consider the rotation from the NED frame to the body frame in which ψ is rotated by
R , added to θ , and this sum then rotated by R and finally added to φ . This process is
y ,θ x ,φ
given in (12).
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 153

⎡φ ⎤ ⎛ ⎡0⎤ ⎡0⎤⎞

⎢ ⎥ ⎜ ⎢0⎥⎟
b
ωnb = ⎢ 0 ⎥ + RxT,φ ⎜ ⎢⎢θ ⎥⎥ + RTy ,θ ⎢ ⎥⎟ (12)
⎢0 ⎥ ⎜ ⎢0⎥ ⎟
⎣ ⎦ ⎝⎣ ⎦ ⎣⎢ψ ⎦⎥ ⎠
By expanding (12), the matrix for transforming the rotational velocities from the NED frame
to the body frame is defined as (13),

⎡1 0 − sin θ ⎤
⎢ ⎥
TΘ−1 ( Θ ) := ⎢0 cos φ cosθ sin φ ⎥ (13)
⎣⎢0 − sin φ cosθ cos φ ⎦⎥

and therefore the matrix for transforming the rotational velocities from the body frame to
the NED frame is given in (14).

⎡ 1 sin φ tan θ cos φ tan θ ⎤

⎢ ⎥
TΘ ( Θ ) = ⎢0 cos φ − sin φ ⎥ (14)
⎢⎣0 sin φ cosθ cos φ cosθ ⎥⎦

Overall, (1) achieves rotation from the body frame to the NED frame, and by taking the
inverse of (8), rotation from the NED frame to the body frame can be achieved, as shown in
(15).

ν = J −1 (η )η (15)

2.3.3 Kinetic equation

The 6 DoF nonlinear dynamic equations of motion of an underwater vehicle can be
conveniently expressed as (16) (Fossen, 1994, 2002).

Mν + C (ν )ν + D (ν )ν + g (η ) = τ + ω (16)

Here, M denotes the 6 × 6 system inertia matrix containing both rigid body and added
mass, as given by (17).

M = M RB + M A (17)

Similar to (17), the 6 × 6 Coriolis and centripetal forces matrix, including added mass, is
given by (18).

C (ν ) = C RB (ν ) + C A (ν ) (18)

Linear and nonlinear hydrodynamic damping are contained within the 6 × 6 matrix D(ν),
and given by (19).

D (ν ) = D + Dn (ν ) (19)

Here, D contains the linear damping terms, and Dn(ν) contains the nonlinear damping
terms.
154 Autonomous Underwater Vehicles

The 6 × 1 vector of gravitational and buoyancy forces and moments are represented in (16)
by g(η), and determined using (20).

⎡ ( W − B) sin θ ⎤
⎢ ⎥
⎢ − ( W − B ) cos θ sin φ ⎥
⎢ − ( W − B) cosθ cos φ ⎥
⎢ ⎥
⎢
( g b ) g (
g (η ) = ⎢ − y W − y B cosθ cos φ + z W − z B cosθ sin φ ⎥
b
⎥
) (20)

⎢
⎢
( ) (
zg W − zb B sin θ + x g W − xb B cosθ cos φ )
⎥
⎥
⎣ (g b ) g (
⎢ − x W − x B cosθ sin φ − y W − y B sin θ ⎥
b ⎦ )
Here, W is the weight of the vehicle, determined using W=mg where m is the dry mass of the
vehicle and g is the acceleration due to gravity. B is the buoyancy of the vehicle which is due
to how much fluid the vehicle displaces while underwater. This will be determined by the
size and shape of the vehicle. Vectors determining the locations of the centre of gravity and
the centre of buoyancy, relative to the origin of the body frame, are given by (21) and (22)
respectively.

T
rgb = ⎡⎣ x g yg zg ⎤⎦ (21)

rbb = [ xb zb ]
T
yb (22)

The 6 × 1 vector of control input forces is denoted by τ, and is given by (23).

τ = [X Y Z K M N ]
T
(23)

Here, the translational forces affecting surge, sway and heave are X, Y, and Z respectively,
and the rotational moments affecting roll, pitch and yaw are K, M and N respectively.
The 6 × 1 vector of external disturbances is denoted by ω.
Overall, (16) provides a compact representation for the nonlinear dynamic equations of
motion of an underwater vehicle, formulated in the body frame. By applying the rotations
contained within (8), (16) can be formulated in the NED frame as given in (24).

Mη (η )η + Cη (ν ,η )η + Dη (ν ,η )η + gη (η ) = τη (η ) + ω (24)

Within (24), the equations in (25) contain the rotations of the various matrices from the body
frame to the NED frame.

Mη (η ) = J −T (η ) MJ −1 (η )
Cη (ν ,η ) = J −T (η ) ⎡⎣C (ν ) − MJ −1 (η ) J (η ) ⎤⎦ J −1 (η )
Dη (ν ,η ) = J −T (η ) D (ν ) J −1 (η ) (25)
gη (η ) = J (η ) g (η )
−T

τη (η ) = J −T (η )τ
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 155

The presence of nonlinearities contained within Dn(ν), combined with the coupling effect of
any non-zero off-diagonal elements within all matrices, can lead to a highly complex model
containing a large number of coefficients.

2.4 Control laws

Various control strategies, and therefore control laws, have been implemented for AUV
systems.
The benchmark for control systems would be the classical proportional-integral-derivative
(PID) control that has been used successfully to control many different plants, including
autonomous vehicles. PID schemes are, however, not very effective in handling the
nonlinear AUV dynamics with uncertain models operating in unknown environments with
strong wave and current disturbances. PID schemes are therefore only generally used for
very simple AUVs working in environments without any external disturbances.
An alternative control scheme known as sliding mode control (SMC), which is a form of
variable structure control, has proven far more effective and robust at handling nonlinear
dynamics with modelling uncertainties and nonlinear disturbances. SMC is a nonlinear
control strategy which uses a nonlinear switching term to obtain a fast transient response
while still maintaining a good steady-state response. Consequently, SMC has been
successfully applied by many researchers in the AUV community. One of the earliest
applications of using SMC to control underwater vehicles was conducted by Yoerger and
Slotine wherein the authors demonstrated through simulations studies on an ROV model,
the SMC controller’s robustness properties to parametric uncertainties (Yoerger & Slotine,
1985). A multivariable sliding mode controller based on state feedback with decoupled
design for independently controlling velocity, steering and diving of an AUV is presented in
Healey and Lienard (1993). The controller design was successfully implemented on NPS
ARIES AUV as reported in Marco and Healey (2001).

2.4.1 PID control

The fundamentals of PID control is that an error signal is generated that relates the desired
state of the plant to the actual state (26),

e ( t ) = xd ( t ) − x ( t ) (26)

Where e(t) is the error signal, xd(t) is the desired state of the plant, and x(t) is the current state
of the plant, and this error signal is manipulated to introduce a corrective action, denoted
τ(t), to the plant.
PID control is named due to the fact that the three elements that make up the corrective
control signal are: proportional to the error signal by a factor of KP, a scaled factor, KI, of the
integral of the error signal, and a scaled factor, KD, of the derivative of the error signal,
respectively (27).

t
d
τ ( t ) = K P e ( t ) + K I ∫ e ( λ ) dλ + KD e (t ) (27)
0
dt

PID control is best suited to linear plants, yet has also been adopted for use on nonlinear
plants even though it lacks the same level of performance that other control systems possess.
156 Autonomous Underwater Vehicles

However, due to the wide use and acceptance of PID control for use in controlling a wide
variety of both linear and nonlinear plants, it is very much employed as the “gold standard”
that control systems are measured against. An example of a PID-based control strategy
applied to underwater vehicles is given in Jalving (1994).

2.4.2 Sliding mode control

As mentioned previously, sliding mode control is a scheme that makes use of a
discontinuous switching term to counteract the effect of dynamics that were not taken into
account at the design phase of the controller.
To examine how to apply sliding mode control to AUVs, firstly (24) is compacted to the
form of (28),

Mη (η )η + f (η ,η , t ) = τη (η ) (28)

where f (η ,η , t ) contains the nonlinear dynamics, including Coriolis and centripetal forces,
linear and nonlinear damping forces, gravitational and buoyancy forces and moments, and
external disturbances.
If a sliding surface is defined as (29),

s = η + cη (29)
where c is positive, it can be seen that by setting s to zero and solving for η results in η
converging to zero according to (30)

η ( t ) = η ( t0 ) ect0 e −ct (30)

regardless of initial conditions. Therefore, the control problem simplifies to finding a control
law such that (31) holds.

lim s ( t ) = 0 (31)
t →∞

This can be achieved by applying a control law in the form of (32),

τ = −T (η ,η ) sign ( s ) , T (η ,η ) > 0 (32)

with T (η ,η ) being sufficiently large. Thus, it can be seen that the application of (32) will
result in η converging to zero.
If η is now replaced by the difference between the current and desired states of the vehicle, it
can be observed that application of a control law of this form will now allow for a reference
trajectory to be tracked.
Two such variants of SMC are the uncoupled SMC and the coupled SMC.
2.4.2.1 Uncoupled SMC
Within the kinetic equation of an AUV, (16), simplifications can be applied that will reduce
the number of coefficients contained within the various matrices. These simplifications can
be applied due to, for example, symmetries present in the body of the vehicle, placement of
centres of gravity and buoyancy, and assumptions based on the level of effect a particular
coefficient will have on the overall dynamics of the vehicle. Thus, the assumption of body
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 157

symmetries allows reduced level of coupling between the various DoFs. An uncoupled SMC
therefore assumes that no coupling exists between the various DoFs, and that simple
manoeuvring is employed such that it does not excite these coupling dynamics (Fossen,
1994). The effect this has on (16) is to remove all off-diagonal elements within the various
matrices which significantly simplifies the structure of the mathematical model of the
vehicle (Fossen, 1994), and therefore makes implementation of a controller substantially
easier.
2.4.2.2 Coupled SMC
Although the removal of the off-diagonal elements reduces the computational complexity of
the uncoupled SMC, it also causes some limitation to the control performance of AUVs,
particularly those operating in highly dynamic environments and required to execute
complex manoeuvres. Taking these two factors into account, these off-diagonal coupling
terms will have an influence on the overall dynamics of the vehicle, and therefore cannot be
ignored at the design phase of the control law.
Coupled SMC is a new, novel control law that retains more of the coupling coefficients
present in (16) compared to the uncoupled SMC (Kokegei et al., 2008, 2009). Furthermore,
even though it is unconventional to design a controller in this way, the body frame is
selected as the reference frame for this controller. This selection avoids the transformations
employed in (24) and (25) used to rotate the vehicle model from the body frame to the NED
frame although it does require that guidance and navigation data be transformed from the
NED frame to the body frame. By defining the position and orientation error in the NED
frame according to (33),

η = ηˆ − ηd (33)

where η̂ represents an estimate of the current position and orientation provided by the
navigation system, and ηd represents the desired position and orientation provided by the
guidance system, a single rotation is required to transform this error from the NED frame to
the body frame.
In general, desired and current velocity and acceleration data are already represented in the
body frame, and as such, no further rotations are required here for the purposes of
implementing a controller in the body frame.
By comparing the number of rotations required to transform the vehicle model into the NED
frame, as seen in (25), for the uncoupled control scheme with the single rotation required by
the coupled control scheme to transform the guidance and navigation data into the body
frame, it can be seen that the latter has less rotations involved, and is therefore less
computationally demanding.

3. Control allocation
The role of the control law is to generate a generalised force to apply to the vehicle such that a
desired state is approached. This force, τ, for underwater vehicles consists of six components,
one for each DoF, as seen in (23). The control allocation system is responsible for distributing
this desired force amongst all available actuators onboard the vehicle such that this
generalised 6 DoF force is realised. This means that the control allocation module must have
apriori knowledge of the types, specifications, and locations, of all actuators on the vehicle.
158 Autonomous Underwater Vehicles

3.1 Role
The role of the control allocation module is to generate the appropriate signals to the
actuators in order for the generalised force from the control law to be applied to the vehicle.
Since the vehicle under consideration is over-actuated, which means multiple actuators can
apply forces to a particular DoF, the control allocation is responsible for utilising all
available actuators in the most efficient way to apply the desired force to the vehicle. Power
consumption is of particular importance for all autonomous vehicles, as it is a key factor in
determining the total mission duration. The control allocation is therefore responsible for
applying the desired forces to the vehicle, while minimising the power consumed.

3.2 Actuators
The force applied to a vehicle due to the various actuators of a vehicle can be formulated as
(34),

τ = TKu (34)
where, for an AUV operating with 6 DoF with n actuators, T is the actuator configuration
matrix of size 6 × n , K is the diagonal force coefficient matrix of size n × n , and u is the
control input of size n × 1 . Actuators are the physical components that apply the desired
force to the vehicle, and the particular configuration of these actuators will determine the
size and structure of T, K and u, with each column of T, denoted ti, in conjunction with the
corresponding element on the main diagonal of K, representing a different actuator.
A vast array of actuators are available to underwater vehicle designers, the more typical of
which include propellers, control fins and tunnel thrusters, and each has their own
properties that make them desirable for implementation within AUVs. For all the following
actuator descriptions lx defines the offset from the origin of the actuator along the x-axis, ly
defines the offset along the y-axis, and lz defines the offset along the z-axis.

3.2.1 Propellers
Propellers are the most common actuators implemented to provide the main translational
force that drives underwater vehicles. These are typically located at the stern of the vehicle
and apply a force along the longitudinal axis of the vehicle. The structure of ti for a propeller
is given in (35).

T
ti = ⎡⎣1 0 0 0 lz −ly ⎤⎦ (35)

As can be seen from (35), if the propeller is positioned such that there is no y-axis or z-axis
offset, the force produced will be directed entirely along the x-axis of the vehicle, with no
rotational moments produced.

3.2.2 Control surfaces

Control surfaces, or control fins, are actuators that utilise Newton’s Third Law of motion to
apply rotational moments to the vehicle. These surfaces apply a force to the water which
causes a deflection in the water’s motion. Hence, the water must also apply a force to the
control surface. Due to this force being applied at a distance from the centre of gravity of the
vehicle, a rotational moment is produced that acts on the vehicle. The typical configuration
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 159

for control surfaces on an AUV is to have four independently controlled fins arranged in
two pairs orientated horizontally and vertically at the stern of the vehicle. The structure of ti
for the horizontal fins is given in (36),

T
ti = ⎡⎣0 0 1 ly −lx 0 ⎤⎦ (36)

and for vertical fins given in (37).

ti = [ 0 1 0 − l z 0 lx ]
T
(37)

These structures of ti show that horizontal surfaces produce a heave force as well as roll and
pitch moments, while vertical surfaces produce a sway force as well as roll and yaw
moments. What must be considered here is that the force being produced by control
surfaces relies on the vehicle moving relative to the water around it. If the vehicle is
stationary compared to the surrounding water, control surfaces are ineffective. However, if
the vehicle is moving relative to the surrounding water, these actuators are capable of
applying forces and moments to the vehicle while consuming very little power.

3.2.3 Tunnel thrusters

The previously mentioned limitation of control surfaces can be overcome by the use of
tunnel thrusters. These thrusters are usually implemented by being placed in tunnels
transverse to the longitudinal axis of the vehicle. Similar to control surfaces, the typical
arrangement is to position two horizontal tunnel thrusters equidistant fore and aft of the
centre of gravity, and two vertical tunnel thrusters also equidistant fore and aft of the centre
of gravity.
The structure of ti for horizontal thrusters is given in (38),

ti = [ 0 1 0 − l z 0 lx ]
T
(38)

while the structure for vertical thrusters is given in (39).

T
ti = ⎡⎣0 0 1 ly −lx 0 ⎤⎦ (39)

What can be observed here is that horizontal thrusters provide a sway force as well as a roll
and yaw moment, while vertical thrusters provide a heave force as well as roll and pitch
moment. In general, horizontal tunnel thrusters are located such that lz is zero, and vertical
thrusters are located such that ly is zero. The result of this choice is that no roll moment is
produced by these actuators.
The advantage of tunnel thrusters is that forces and moments can be produced even if the
vehicle is stationary with respect to the surrounding water. This greatly increases the
manoeuvrability of the vehicle, as control of the vehicle when moving at low speeds is
possible. However, there are limitations associated with the use of these actuators. Firstly,
these actuators consume more power when activated compared to control surfaces. This is
due to force being produced by the thrusters only when the thruster itself is activated. In
contrast, control surfaces consume power when the deflection angle is altered, but require
very little power to hold the surface in place once the desired angle has been achieved.
160 Autonomous Underwater Vehicles

Secondly, thruster efficiency is reduced when the vehicle is moving. Under certain
conditions, an area of low pressure is produced at the exit of the tunnel, which has the effect
of applying a force to the vehicle in the opposite direction to which the water jet from the
tunnel thruster is attempting to provide. The result is less total force being applied to the
vehicle, and therefore reduced performance when moving at non-zero forward speeds
(Palmer et al., 2009).

3.3 Allocation methods

As previously mentioned when reviewing the various actuators for underwater vehicles, if the
vehicle is stationary, control surfaces are ineffective while tunnel thrusters are very useful.
Conversely, if the vehicle is moving, control surfaces are very efficient in providing force
relative to power consumption compared to tunnel thrusters. The role of the control allocation
is therefore to find a compromise between all actuators that both applies the desired
generalised forces and moments to the vehicle while minimising power consumption. This
balancing act allows the vehicle to maintain the manoeuvrability provided by all actuators,
while at the same time allowing for as long a mission duration as possible.

3.3.1 Non-optimal scheme

One of the most straightforward methods for control allocation is application of the inverse
of (34), i.e., (40) (Fossen, 2002).

−1
u = (TK ) τ (40)

This method is very simple to implement as it consists of a single matrix multiplication.

Therefore it is easy and efficient to implement within the computational processing
constraints of an AUV. However, due to its simplicity, no attempt is made to minimise
power consumption. If the vehicle contains both control surfaces and tunnel thrusters, both
of these types of actuators will be utilised equally, even when the vehicle is moving at
maximum velocity with respect to the surrounding fluid. However, since it is much more
efficient to utilise control surfaces rather than tunnel thrusters while the vehicle is moving
with respect to the water, a more intelligent approach is desired for implementing control
allocation to minimise power consumption.

3.3.2 Quadratic programming

The limitations of the aforementioned non-optimal scheme can be overcome by formulating a
quadratic programming optimisation problem to solve for actuator inputs (Fossen, 2002,
Fossen et al., 2009). By introducing a weighting matrix into the problem statement, actuator
usage can be biased towards utilising control surfaces over tunnel thrusters. Therefore, the
generalised force desired from the control law can be realised by the actuators while
minimising power consumption. There are however, limitations associated with this scheme.
Firstly, although it is possible to calculate an explicit solution to this problem, in the event of
actuator reconfiguration, such as an actuator failure, this explicit solution would need to be
recalculated, which can be computationally intensive. Iterative approaches, such as sequential
quadratic programming, can be implemented that allow for actuator failures, but this method
has the potential to require several iterations of the programming problem be solved at each
control sample interval. Again, this can be a computationally intensive task.
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 161

3.3.3 2-Stage scheme

A third scheme that is proposed here for the implementation of the control allocation is to
break the control allocation problem into two smaller sub-problems, as seen in Figure 2,
where the first sub-problem addresses control allocation to the main propeller and control
surfaces, and the second addresses control allocation to the tunnel thrusters. Using this type
of scheme, the control surfaces can firstly be used to their full extent in order to realise the
generalised force as closely as possible. Only after full utilisation of the control surfaces
occurs will the tunnel thrusters be introduced to provide forces and moments that the
surfaces alone cannot produce. Using this methodology, the low power consumption
control surfaces will be used as much as possible, while the higher power consumption
tunnel thrusters will be called upon only when required to provide the extra manoeuvring
capabilities that they possess. Furthermore, in a situation where accurate manoeuvring is
not required, such as traversing from one waypoint to the next with no concern for what
trajectory the vehicle follows, the second stage can be disabled such that no tunnel thruster
is used, and control is performed entirely by the main propeller and control surfaces.
However, if trajectory tracking is desired, the thruster allocation module can still be enabled.
This will allow for the situation when the control surfaces provide inadequate force to the
vehicle, and therefore the thrusters can assist in providing the extra force required to
maintain the vehicle tracking the desired trajectory.

Fig. 2. 2-Stage Control Allocation Scheme Block Diagram.

Implementation of this scheme would look somewhat like a 2-stage non-optimal scheme, as
seen in Figure 2. The first stage would require the matrix operation (TK)-1τ for the main
propeller and control surfaces in order to obtain as much force required from these
actuators. An estimate of the force produced for this particular set of control values would
then be calculated such that this force estimate can be subtracted from the total force
required. Any residual force requirement would then become the input to the second stage
of the control allocation, which would perform the matrix operation (TK ) τ for the tunnel
−1

thrusters in order for these actuators to provide any extra force that the control surfaces
cannot deliver alone. Therefore, the computational requirement for this scheme is quite
minimal compared to the quadratic programming scheme, yet still heavily biases the use of
control surfaces over tunnel thrusters.
162 Autonomous Underwater Vehicles

4. Overall control system

Overall, the control system of any autonomous underwater vehicle consists of a control law
module and a control allocation module. The former is responsible for generating the
generalised force in 6 DoF based on current and desired states, while the latter is responsible
for distributing this generalised force amongst the actuators of the vehicle. Due to the
modularity of such a control system, each of these subsystems can be designed and
implemented independently, yet both must work cooperatively in order to accurately
control the vehicle.
As discussed, the inputs to the control system are the desired state from the guidance
system, and the current state from the navigation system. These are also the inputs to the
control law. The outputs from the control system are the control signals that are sent to the
individual actuators on the vehicle, which are the outputs generated by the control
allocation. The intermediate signals connecting the control law to the control allocation are
the 6 DoF generalised forces which are the outputs from the control law and therefore the
inputs to the control allocation. Figure 3 shows the overall control system. As can be seen,
the modularity of the system allows for control laws and control allocation schemes to be
easily replaced, provided the inputs and outputs are of the same dimensions.

xd Control Law τ Control u

Allocation

Fig. 3. Overall Control System Block Diagram.

For comparison purposes, both the uncoupled SMC law and the coupled SMC law are
simulated here. Both simulations consist of identical control allocation modules that contain
the previously described 2-stage scheme. The vehicle model used for these simulations is
based on that which was proposed by Prestero (2001a, 2001b). This is a mathematical model
for the REMUS (Remote Environmental Monitoring Unit) underwater vehicle developed by
the Woods Hole Oceanographic Institute’s Oceanographic Systems Laboratory. Alterations
have been made to this model, based on current technology, which enable the vehicle to
become fully actuated. These changes include the placement of four tunnel thrusters, two
fore and two aft, such that sway, heave, pitch, and yaw motions are possible without any
water flow over the control surfaces. Furthermore, the position of the vertical thrusters
along the y-axis of the body, and the position of the horizontal thrusters along the z-axis of
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 163

the body are set to 0m. Hence, there is no roll moment applied to the vehicle when these
thrusters are activated. The thrusters introduced to the model are based on the 70mm
IntegratedThrusterTM produced by TSL Technology Ltd. This particular device can provide
a maximum thrust of 42N, and due to its compact size, is well suited to this particular
application. Also, the propulsion unit has been altered within the model. The simulation
model used contains a propulsion unit based on the Tecnadyne Model 540 thruster. This
device is able to provide approximately 93.2N of thrust and, with a propeller diameter of
15.2cm, is therefore well suited for this application.
The trajectory that the vehicle is asked to follow here consists of a series of unit step
inputs applied to each DoF. The translational DoFs experience a step input of 1 metre
whereas the rotational DoFs experience a step input of 1 radian, or approximately 60°. All
inputs are applied for a period of 20 seconds, such that both transient and steady state
behaviour can be observed. These unit step inputs excite the vehicle in all combinations of
DoFs, from a single DoF, through to all DoFs being excited at once. This simulation
assumes no water current, and therefore no water flow over the vehicle when it is
stationary. Hence, roll cannot be compensated for when stationary and for this reason this
simulation does not excite the roll component of the model. All other DoFs are excited,
however.
Due to the way the aforementioned trajectory is supplied to the vehicle model, no guidance
system is present. The step inputs will be applied at set times regardless of the state of the
vehicle, and therefore observation of the control performance can be observed without
influence from unnecessary systems. Therefore total execution time is constant for all
simulations. Performance metrics used here in evaluating each control system is the
accumulated absolute error between the desired translation/rotation and actual
translation/rotation for each DoF. By looking at the following plots, observations can be
made regarding such time-domain properties as rise time, settling time and percentage
overshoot.
Figures 4-13 show the desired and actual trajectories for each individual DoF when this
complex set of manoeuvres is applied to both the uncoupled SMC and the coupled SMC.
Figures 4, 6, 8, 10 and 12 show the complete trajectory for each DoF, while Figures 5, 7, 9, 11
and 13 show these same trajectories with the focus being on the last 300 seconds of the
mission. This latter section of the mission is when multiple DoFs, particularly the rotational
DoFs, are excited simultaneously.
Figure 4 shows the surge motion of the vehicle for the two different control systems. As can
be seen, both systems exhibit desired properties for the first 180 seconds. During this period,
all manoeuvring is exciting only the translational DoFs which the main propeller and tunnel
thrusters can handle independently. After 180 seconds, the other DoFs are also excited, and
the effect of this combined motion produces significant overshoots, especially for the
uncoupled system, within the surge motion of the vehicle. This is more easily seen in Figure
5 where the larger overshoots can be seen for the uncoupled system compared to the
coupled system.
The sway motion of the AUV is shown in Figure 6. Minor overshooting is observed for both
systems, especially when rotational DoFs are excited in combination with the translational
DoFs. This can be observed in Figure 7. However, when the sway motion is excited,
convergence to the desired set point is observed.
164 Autonomous Underwater Vehicles

Fig. 4. Surge trajectory.

Fig. 5. Surge trajectory magnified.

Heave motion is shown in Figure 8. Here, the initial position of the vehicle is chosen to be 10
metres below the surface of the water, such that the vehicle is completely submerged for the
entire trajectory. When observing the heave motion in Figure 9, it is clear that when heave is
excited, convergence to the desired set point is achieved. However, it is also evident that
significant coupling exists between this DoF and other DoFs as divergence from the desired
set point occurs when the heave motion is not excited.
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 165

Fig. 6. Sway trajectory.

Fig. 7. Sway trajectory magnified.

Pitch motion is observed in Figure 10. The first observation that can be made from this plot
is the relatively large spikes in the motion for both the coupled and uncoupled systems
between 800 seconds and 1000 seconds. The second observation is that a steady state error is
observed when the uncoupled system attempts a pitch of 1 radian. This is more clearly
observed in Figure 11. Due to the vertical offset of the vehicle’s centre of buoyancy from its
centre of gravity, a restorative moment will always be applied to the vehicle for non-zero
pitch angles. This plot indicates that it is difficult for the uncoupled system to compensate
for this effect. However, the coupled system is able to not only eliminate this steady state
error, it also has significantly smaller overshoot in general, as can be seen in Figure 10.
166 Autonomous Underwater Vehicles

Fig. 8. Heave trajectory.

Fig. 9. Heave trajectory magnified.

Figure 12 shows the desired yaw motion of the vehicle. Here we can see significant
overshooting from both systems, especially during the time period before the yaw motion is
excited. By observing Figure 13, it can be seen that the coupled system achieves faster
dynamics in terms of rise time, but the cost of this is overshoot. The uncoupled system has
significantly less overshoot, and this is due to it taking slightly longer to react to trajectory
changes.
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 167

Fig. 10. Pitch trajectory.

Fig. 11. Pitch trajectory magnified.

Overall, the simulation trajectory used here is extremely complex as it excites all possible
combinations of DoFs without any water current present. This level of complexity allows for
the coupling that exists between DoFs within an AUV model to be highlighted. For example,
looking at the period of 0 to 660 seconds in Figure 12, the yaw motion is not being excited,
yet motion in this DoF is observed. This coupling between multiple DoFs is the reason why
controlling AUVs is a complex and challenging task.
168 Autonomous Underwater Vehicles

Fig. 12. Yaw trajectory.

Fig. 13. Yaw trajectory magnified.

Furthermore, if the absolute error between actual and desired position is accumulated over
the entire mission for each DoF, a measure of the accuracy of the control system can be
obtained. A comparison can be made using Table 1. Here, we can see that by summing
together the values obtained from the three translational DoFs, the coupled system
improves upon the uncoupled system by 9.75%, while by summing together the values
obtained from the three rotational DoFs, the coupled improves upon the uncoupled system
by 23.87%. This indicates that the coupled system is superior to the uncoupled system for
both translational and rotational motion, when looking at the accumulated absolute error.
Fully Coupled 6 Degree-of-Freedom Control of an Over-Actuated Autonomous Underwater Vehicle 169

Accumulated Absolute Accumulated Absolute

Control System
Translational Error Rotational Error
Uncoupled System 1.3173x104 (metres) 8.3713x103 (radians)
Coupled System 1.2003x104 (metres) 6.7580x103 (radians)
Table 1. Accumulated Absolute Translational and Rotational Errors.

5. Conclusions
Due to the increased adoption of AUVs for civilian and defence operations, accuracy and
reliability are two key factors that enable an AUV to successfully complete its mission. The
control system is just one of the various components within the autonomy architecture of an
AUV that helps in achieving this goal. Within the control system, the control law should be
robust to both external disturbances and model parameter uncertainties, while the control
allocation should utilise the various actuators of the vehicle to apply the desired forces to
the vehicle while minimising the power expended.
PID control has been successfully implemented on a variety of systems to effectively
provide compensation. However, since PID control is better suited to linear models, the
level of performance provided by PID control is not to the same standard as other,
particularly nonlinear, control schemes when applied to complex nonlinear systems. Sliding
mode control has proven to be a control law that is robust to parameter uncertainties, and
therefore is a prime candidate for implementation within this context due to the highly
complex coupled nonlinear underwater vehicle model. Active utilisation of the coupled
structure of this model is what coupled SMC attempts to achieve, such that induced motion
in one DoF due to motion in another DoF is adequately compensated for. This is where
coupled SMC has a distinct advantage over uncoupled SMC for trajectory tracking
applications when multiple DoFs are excited at once.
Various schemes exist for control allocation with the ultimate goal being to apply the
desired generalised forces while minimising power consumption, both due to the actuator
usage and computational demands. Non-optimal schemes exist where a generalised inverse
of the force produced by all actuators is used as the allocation scheme, with the limitation
being that there is no functionality to bias actuators under certain operating conditions, such
as utilising control surfaces over thrusters during relatively high speed manoeuvring.
Quadratic programming incorporates a weighting matrix that can bias control surface usage
over tunnel thrusters, and has been implemented both online and offline, with each having
advantages and disadvantages. Online optimisation allows for changes to the actuator
configuration, such as failures or varied saturation limits, but is computationally
demanding. Offline optimisation is less computationally demanding during mission
execution, but cannot allow for altered actuator dynamics. A compromise between these
schemes is the proposed 2-stage scheme where control surfaces are utilised to their full
extent, and the tunnel thrusters used only when needed.
Overall, the goal of the control system is to provide adequate compensation to the vehicle,
even in the presence of unknown and unmodelled uncertainties while also minimising
power consumption and therefore extending mission duration. Choosing wisely both the
control law and the control allocation scheme within the overall control system is
fundamental to achieving both of these goals.
170 Autonomous Underwater Vehicles

6. Acknowledgements
The financial support of this research from the Australian Government’s Flagship
Collaboration Fund through the CSIRO Wealth from Oceans Flagship Cluster on Subsea
Pipelines is acknowledged and appreciated.

7. References
Fossen, T. I. (1994). Guidance and Control of Ocean Vehicles, John Wiley & Sons, Inc., ISBN 0-
471-94113-1, Chichester, England
Fossen, T. I. (2002). Marine Control Systems: Guidance, Navigation and Control of Ships, Rigs and
Underwater Vehicles (1), Marine Cybernetics, ISBN 82-92356-00-2, Trondheim, Norway
Fossen, T. I., Johansen, T. A. & Perez, T. (2009). A Survey of Control Allocation Methods for
Underwater Vehicles, In: Underwater Vehicles, Inzartsev, A. V., pp. (109-128), In-
Tech, Retrieved from
<http://www.intechopen.com/articles/show/title/a_survey_of_control_allocatio
n_methods_for_underwater_vehicles>
Healey, A. J. & Lienard, D. (1993). Multivariable Sliding Mode Control for Autonomous
Diving and Steering of Unmanned Underwater Vehicles. IEEE Journal of Oceanic
Engineering, Vol. 18, No. 3, (July), pp. (327-339), ISSN 0364-9059
Jalving, B. (1994). The NDRE-AUV Flight Control System. IEEE Journal of Oceanic
Engineering, Vol. 19, No. 4, (October), pp. (497-501), ISSN 0364-9059
Kokegei, M., He, F. & Sammut, K. (2008). Fully Coupled 6 Degrees-of-Freedom Control of
Autonomous Underwater Vehicles, MTS/IEEE Oceans '08, Quebec City, Canada,
September 15-18
Kokegei, M., He, F. & Sammut, K. (2009), Nonlinear Fully-Coupled Control of AUVs, Society
of Underwater Technology Annual Conference, Perth, Australia, 17-19 February
Lammas, A., Sammut, K. & He, F. (2010). 6-DoF Navigation Systems for Autonomous
Underwater Vehicles, In: Mobile Robots Navigation, Barrera, A., pp. (457-483), In-Teh,
Retrieved from <http://www.intechopen.com/articles/show/title/6-dof-
navigation-systems-for-autonomous-underwater-vehicles>
Lammas, A., Sammut, K. & He, F. (2008), Improving Navigational Accuracy for AUVs using the
MAPR Particle Filter, MTS/IEEE Oceans '08, Quebec City, Canada, 15-18 September
Marco, D. B. & Healey, A. J. (2001). Command, Control, and Navigation Experimental
Results with the NPS ARIES AUV. IEEE Journal of Oceanic Engineering, Vol. 26, No.
4, (October), pp. (466-476), ISSN 0364-9059
Palmer, A., Hearn, G. E. & Stevenson, P. (2009). Experimental Testing of an Autonomous
Underwater Vehicle with Tunnel Thrusters, First International Symposium on Marine
Propulsors, Trondheim, Norway, 22-24 June
Prestero, T. (2001a), Development of a Six-Degree of Freedom Simulation Model for the
REMUS Autonomous Underwater Vehicle, 12th International Symposium on
Unmanned Untethered Submersible Technology, University of New Hampshire,
Durham, NH, 26-29 August
Prestero, T. (2001b). Verification of a Six-Degree of Freedom Simulation Model for the REMUS
Autonomous Underwater Vehicle. Master of Science in Ocean Engineering and Master
of Science in Mechanical Engineering, Massachusetts Institute of Technology and
the Woods Hole Oceanographic Institution, Cambridge and Woods Hole
Yoerger, D. R. & Slotine, J.-J. E. (1985). Robust Trajectory Control of Underwater Vehicles. IEEE
Journal of Oceanic Engineering, Vol. 10, No. 4, (October), pp. (462-470), ISSN 0364-9059
 Part 3

Mission Planning and Analysis

 0
8

Short-Range Underwater Acoustic

Communication Networks
Gunilla Burrowes and Jamil Y. Khan
The University of Newcastle
Australia

1. Introduction
This chapter discusses the development of a short range acoustic communication channel
model and its properties for the design and evaluation of MAC (Medium Access Control)
and routing protocols, to support network enabled Autonomous Underwater Vehicles (AUV).
The growth of underwater operations has required data communication between various
heterogeneous underwater and surface based communication nodes. AUVs are one such
node, however, in the future, AUV’s will be expected to be deployed in a swarm fashion
operating as an ad-hoc sensor network. In this case, the swarm network itself will be
developed with homogeneous nodes, that is each being identical, as shown in Figure 1, with
the swarm network then interfacing with other fixed underwater communication nodes. The
focus of this chapter is on the reliable data communication between AUVs that is essential to
exploit the collective behaviour of a swarm network.
A simple 2-dimensional (2D) topology, as shown in Figure 1(b), will be used to investigated
swarm based operations of AUVs. The vehicles within the swarm will move together, in a
decentralised, self organising, ad-hoc network with all vehicles hovering at the same depth.
Figure 1(b) shows the vehicles arranged in a 2D horizontal pattern above the ocean floor

AUV 3

AUV 2

AUV 1

AUV 6

AUV 5

AUV 4

AUV 9

AUV 8
--- Communication Path 1
AUV 7

Inter-node Range (m) --- Communication Path 2

Depth (m)

(a) AUV Swarm demonstrating stylised (b) 2D AUV Swarm Topology

SeaVision©vehicles

Fig. 1. Swarm Architecture

174
2 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

giving the swarm the maximum coverage area at a single depth, while forming a multi-hop
communication network. The coverage area will depend on application. For example, the
exploration of oil and gas deposits underwater using hydrocarbon sensing would initially
require a broad structure scanning a large ocean footprint before narrowing the range between
vehicles as the sensing begins to target an area. Thus vehicles may need to work as closely
as 10 m with inter-node communication distance extending out to 500 m. These operating
distances are substantially shorter than the more traditional operations of submarines and
underwater sensor to surface nodes that have generally operated at greater than 1km. Thus,
the modelling and equipment development for the communication needs of these operations
has focused on longer range data transmission and channel modelling. To exploit the full
benefits of short range communication systems it is necessary to study the properties of short
range communication channels.
Most AUV development work has concentrated on the vehicles themselves and their
operations as a single unit (Dunbabin et al., 2005; Holmes et al., 2005), without giving
much attention to the development of the swarm architecture which requires wireless
communication networking infrastructure. To develop swarm architectures it is necessary to
research effective communication and networking techniques in an underwater environment.
Swarm operation has many benefits over single vehicle use. The ability to scan or ’sense’
a wider area and to work collaboratively has the potential to vastly improve the efficiency
and effectiveness of mission operations. Collaboration within the swarm structure will
facilitate improved operations by building on the ability to operate as a team which will result
in emergent behaviours that are not exhibited by individual vehicles. A swarm working
collaboratively can also help to mitigate the problem of high propagation delay and lack
of bandwidth available in underwater communication environments. Swarm topology will
facilitate improved communication performance by utilising the inherent spatial diversity
that exists in a large structure. For example, information can be transmitted more reliably
within a swarm architecture by using multi-hop networking techniques. In such cases, loss
of an individual AUV, which can be expected at times in the unforgiving ocean environment,
will have less detrimental effect compared to a structure where multiple vehicles operate on
their own. (Stojanovic, 2008).
The underwater acoustic communication channel is recognised as one of the harshest
environments for data communication, with long range calculations of optimal channel
capacity of less than 50kbps for SNR (Signal-to-Nosie Ratio) of 20dB (Stojanovic, 2006) with
current modem capacities of less than 10kbps (Walree, 2007). Predictability of the channel
is very difficult with the conditions constantly changing due to seasons, weather, and the
physical surroundings of sea floor, depth, salinity and temperature. Therefore, it must be
recognised that any channel model needs to be adaptable so that the model can simulate the
channel dynamics to be able to fully analyse the performance of underwater networks.
In general, the performance of an acoustic communication system underwater is characterised
by various losses that are both range and frequency dependent, background noise that is
frequency dependent and bandwidth and transmitter power that are both range dependent.
The constraints imposed on the performance of a communication system when using an
acoustic channel are the high latency due to the slow speed of the acoustic signal propagation,
at 0.67 ms/m (compared with RF (Radio Frequency) in air at 3.3 ns/m), and the signal fading
properties due to absorption and multipath. Specific constraints on the performance due to
the mobility of AUV swarms is the Doppler effect resulting from any relative motion between
Short-Range Underwater Acoustic Communication Networks 1753
Short-Range Underwater Acoustic
Communication Networks

a transmitter and a receiver, including any natural motion present in the oceans from waves,
currents and tides.
Noise in the ocean is frequency dependent. There are three major contributors to noise
underwater: ambient noise which represents the noise in the far field; self noise of the vehicle
(considered out of band noise); and intermittent noise sources including noises from biological
sources such as snapping shrimp, ice cracking and rain. Ambient noise is therefore the
component of noise taken into account in acoustic communication performance calculations.
It is characterised as a Gaussian Distribution but it is not white as it does not display a
constant power spectral density. For the frequencies of interest for underwater acoustic
data communication, from 10 to 100 kHz, the ambient noise value decreases with increasing
frequency. Therefore, using higher signal frequencies, which show potential for use in shorter
range communication, will be less vulnerable to the impact of ambient noise.
Short range underwater communication systems have two key advantages over longer range
operations; a lower end-to-end delay and a lower signal attenuation. End-to-end propagation
at 500 m for example is approximately 0.3 sec which is considerable lower than the 2 sec at
3 km but still critical as a design parameter for shorter range underwater MAC protocols.
The lower signal attenuation means potentially lower transmitter power requirements which
will result in reduced energy consumption which is critical for AUVs that rely on battery
power. Battery recharge or replacement during a mission is difficult and costly. The dynamics
associated with attenuation also changes at short range where the spreading component
dominates over the absorption component, which means less dependency on temperature,
salinity and depth (pressure). This also signifies less emphasis on frequency as the frequency
dependent part of attenuation is in the absorption component and thus will allow the use
of higher signal frequencies and higher bandwidths at short ranges. This potential needs
to be exploited to significantly improve the performance of an underwater swarm network
communication system.
A significant challenge for data transmission underwater is multipath fading. The effect of
multipath fading depends on channel geometry and the presence of various objects in the
propagation channel. Multipath’s occur due to reflections (predominately in shallow water),
refractions and acoustic ducting (deep water channels), which create a number of additional
propagation paths, and depending on their relative strengths and delay values can impact on
the error rates at the receiver. The bit error is generated as a result of inter symbol interference
(ISI) caused by these multipath signals. For very short range single transmitter-receiver
systems, there could be some minimisation of multipath signals (Hajenko & Benson, 2010;
Waite, 2005). For swarm operations, however, there is potentially a different mix of multipath
signals that need to be taken into account, in particular, those generated due to the other
vehicles in the swarm.
Careful consideration of the physical layer parameters and their appropriate design will help
maximise the advantages of a short range communications system that needs to utilise the
limited resources available in an underwater acoustic networking environment.
The following section will introduce the parameters associated with acoustic data
transmission underwater. The underwater data transmission channel characteristics will be
presented in Section 3 with a discussion of the advantages and disadvantages of the short
range channel. Section 4 will show how these will impact on AUV swarm communications
and the development of a short range channel model for the design and evaluation of MAC
and routing protocols. This is followed in Section 5 by a discussion of the protocol techniques
required for AUV swarm network design.
176
4 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

2. Introduction to acoustic underwater communication network

The underwater data communication link and networking environment presents a
substantially different channel to the RF data communication channel in the terrestrial
atmosphere. Figure 2 illustrates a typical underwater environment for data transmission using
a single transmitter-receiver pair.

Fig. 2. Underwater Acoustic Environment

A simple schematic of the data transmission scheme involving a projector (transmitter) and a
hydrophone (receiver) is presented in Figure 3. The projector takes the collected sensor and
navigational data and formats it into packets at the Data Source and this is then modulated
with the carrier frequency. The modulated signal is amplified to a level sufficient for signal
reception at the receiver. There is an optimum amplification level as there is a trade-off
between error free transmission and conservation of battery energy. The acoustic power
radiated from the projector as a ratio to the electrical power supplied to it, is the efficiency ηtx
of the projector and represented by the Electrical to Acoustic conversion block. On the receiver
side, the sensitivity of the hydrophone converts the sound pressure that hits the hydrophone
to electrical energy, calculated in dB/V. Signal detection, includes amplification and shaping
of the input to determine a discernible signal. Here a detection threshold needs to be reached
and is evaluated as the ratio of the mean signal power to mean noise power (SNR). The carrier
frequency is then supplied for demodulation, before the transmitted data is available for use
within the vehicle for either data storage or for input into the vehicles control and navigation
requirements.
Underwater data communication links generally support low data rates mainly due to the
constraints of the communication channel. The main constraints are the high propagation
delay, lower effective SNR and lower bandwidth. The effects of these constraints could be
reduced by using short distance links and the use of multi-hop communication techniques to
Short-Range Underwater Acoustic Communication Networks 1775
Short-Range Underwater Acoustic
Communication Networks

cover longer transmission ranges. For an AUV swarm network, use of the above techniques
could be crucial to design an effective underwater network. To develop a multi-node
swarm network it is necessary to manage all point to point links using a medium access
control (MAC) protocol. In a multi-access communication system like a swarm network a
transmission channel is shared by many transceivers in an orderly fashion to transmit data
in an interference free mode. Figure 2 shows a point to point communication link with two
AUVs. When a network is scaled up to support N number of AUVs then it becomes necessary
to control multiple point to point or point to multi-point links.

PROJECTOR HYDROPHONE

Receiver
Receiver
Transmitter

Data Modula Power Electrical Acoustic Signal Demodu Data

Source tion Amplifier to to Detection lation Storage/
Acoustic Electrical and Reuse
Receiver conversion conversion Amplification

Carrier Carrier
Signal Signal

Range (m)

Fig. 3. Block Diagram of Projector and Hydrophone

To control the transmission of data it is necessary to design an effective MAC protocol which
can control transmission of information from different AUVs. The design of a MAC protocol
in a swarm network could be more complex if a multi-hop communication technique is
used. The multi-hop communication technique will allow a scalable network design as well
as it can support long distance transmission without the need of high power transmitter
and receiver circuits. For example, using a multi-hop communication technique if AUV3
in Figure 1(b) wants to transmit packets to AUV7 then it can potentially use a number
of communication paths to transmit packets. Some of the possible paths from AUV3 to
AUV7 are: AUV3-AUV2-AUV1-AUV4-AUV7 or AUV3-AUV6-AUV9-AUV8-AUV7. The path
selection in a network is controlled by the routing protocols. Optimum routing protocols
generally select transmission paths based on a number of factors. However, the main
factor used to select an optimum path in a wireless network is the SNR which indicates
the quality of a link. Similarly the MAC protocol will use the transmission channel state
information to develop an optimum packet access technique. To effectively design these
protocols it is necessary to understand the properties of short range underwater channel
characteristics. Before moving into the protocol design issues we will first evaluate the short
range underwater channel characteristics in the following Sections.

3. Underwater data transmission channel characteristics

This section will focus on the parameters of the ocean channel that will affect the acoustic
signal propagation from the projector to the hydrophone. There are well established
178
6 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

underwater channel models that will be used to derive and present the data transmission
characteristics for a short-range link.

3.1 Acoustic signal level

The projector source level, SLtprojector , is generally defined in terms of the sound pressure
level at a reference distance of 1 m from its acoustic centre. The source intensity at this
reference range is I = Ptx /Area (W/m2 ) and measured in dB ’re 1 μPa’ but strictly meaning
’re the intensity due to a pressure of 1 μPa’. For an omni directional projector the surface
area is a sphere (4πr2 = 12.6m2 ). Thus, SL projector = 10log(( Ptx /12.6)/Ire f ) dB, where Ptx
is the total acoustic power consumed by projector and the reference wave has an intensity:
Ire f = ( Pare f )2 /ρ ∗ c (Wm−2 ) where reference pressure level; Pare f is 1 μPa, ρ is the density
of the medium and; c is the speed of sound (averages for sea water: ρ = 1025 kg/m3 and
c=1500 m/s) (Coates, 1989; Urick, 1967).
The equation for the transmitter acoustic signal level (SL projector ) at 1 m for an
omni-directional projector can be written:

SL projector ( P) = 170.8 + 10logPtx dB (1)

Idir
If the projector is directional, then the projector directivity index is DItx = 10log( Iomni ) where
Iomni is the intensity if spread spherically and Idir is the intensity along the axis of the beam
pattern. Directivity can increase the source level by 20dB (Waite, 2005). The more general
equation for the transmitter acoustic signal level (SL projector ) can be written:

SL projector ( P, η, DI ) = 170.8 + 10logPtx + 10logηtx + DItx dB (2)

where the efficiency of the projector ηtx takes into account the losses associated with the
electrical to acoustic conversion as shown in Figure 3, thus reducing the actual SL radiated
by the projector. This efficiency is bandwidth dependent and can vary from 0.2 to 0.7 for a
tuned projector (Waite, 2005).

3.2 Signal attenuation

Sound propagation in the ocean is influenced by the physical and chemical properties of
seawater and by the geometry of the channel itself. An acoustic signal underwater experiences
attenuation due to spreading and absorption. In addition, depending on channel geometry
multipath fading may be experienced at the hydrophone. Path loss is the measure of the lost
signal intensity from projector to hydrophone. Understanding and establishing a accurate
path loss model is critical to the calculations of Signal-to-Noise ratio (SNR).

3.2.1 Spreading loss

Spreading loss is due to the expanding area that the sound signal encompasses as it
geometrically spreads outward from the source.

PLspreading (r ) = k ∗ 10log(r ) dB (3)

where r is the range in meters and k is the spreading factor.

When the medium in which signal transmission occurs is unbounded, the spreading is
spherical and the spreading factor k=2 whereas in bounded spreading, considered as
cylindrical k=1. Urick (1967) suggested that spherical spreading was a rare occurrence in
the ocean but recognised it may exist at short ranges. As AUV swarm operations will occur
Short-Range Underwater Acoustic Communication Networks 1797
Short-Range Underwater Acoustic
Communication Networks

at short range it is likely that spherical spreading will need to be considered which means
a higher attenuation value. Spreading loss is a logarithmic relationship with range and its
impacts on the signal is most significant at very short range up to approximately 50m as
seen in Figure 5(a). At these shorter ranges spreading loss plays a proportionally larger part
compared with the absorption term (which has a linear relationship with range).

3.2.2 Absorption loss

The absorption loss is a representation of the energy loss in the form of heat due to the
viscous friction and ionic relaxation that occurs as the wave generated by an acoustic signal
propagates outwards and this loss varies linearly with range as follows:

PL absorption (r, f ) = 10log(α( f )) ∗ r dB (4)

where r is range in kilometres and α is the absorption coefficient.
More specifically the absorption of sound in seawater is caused by three dominant effects;
viscosity (shear and volume) , ionic relaxation of boric acid and magnesium sulphate
( MgSO4 ) molecules and the relaxation time. The effect of viscosity is significant at high
frequencies above 100 kHz, whereas the ionic relaxation effects of magnesium affect the mid
frequency range from 10 kHz up to 100 kHz and boric acid at low frequencies up to a few kHz.
In general, the absorption coefficient, α, increases with increasing frequency and decreases
as depth increases (Domingo, 2008; Sehgal et al., 2009) and is significantly higher in the sea
compared with fresh water due predominately to the ionic relaxation factor.
Extensive measurements of absorption losses over the last half century has lead to several
empirical formulae which take into account frequency, salinity, temperature, pH, depth and
speed of sound. A popular version is Thorp’s expression (Thorp, 1965), Equation 5, which is
based on his initial investigations in the 60’s and has since been converted into metric units
(shown here). It is valid for frequencies from 100Hz to 1MHz and is based on seawater with
salinity of 35% ppt, pH of 8, temp of 4◦ C and depth of 0 m (atmospheric pressure) which is
assumed but not stated by Thorp.

0.11 f 2 44 f 2
α( f ) = + + 2.75 × 10−4 f 2 + 0.0033 dB/km (5)
1+ f 2 4100 + f 2

Fisher and Simmons (1977) and others (Francois & Garrison, 1982) have since proposed
other variations of α. In particular, Fisher and Simmons in the late 70’s found the effect
associated with the relaxation of boric acid on absorption and provided a more detailed
form of absorption coefficient α in dB/km which varies with frequency, pressure (depth) and
temperature (also valid for 100 Hz to 1 MHz with salinity 35% ppt and acidity 8 pH)(Fisher &
Simmons, 1977; Sehgal et al., 2009), given in Equation 6.

A1 f 1 f 2 A P f f2
α( f , d, t) = + 22 2 2 2 + A3 P3 f 2 dB/km (6)
f1 + f
2 2 f2 + f

where d is depth in meters and t is temperature in ◦ C. The ’A’ coefficients represent the effects
of temperature, while the ’P’ coefficients represent ocean depth (pressure) and f 1 , f 2 represent
the relaxation frequencies of Boric acid and ( MgSO4 ) molecules. These terms were developed
by Fisher and Simmons (1977) and presented more recently by (Domingo, 2008; Sehgal et al.,
2009).
180
8 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

Fig. 4. Absorption Coefficient vs Frequency

Figure 4 shows the absorption coefficients in dB/km vs signal frequency for both Thorp
and Fisher and Simmons coefficients and shows that in general α increases with increasing
frequency at any fixed temperature and depth. Up until around 80kHz temperature change
has a more significant affect on α than depth (Waite, 2005), but above these frequencies depth
begins to dominate (Domingo, 2008; Sehgal et al., 2009). In any case, Thorps ’approximation’ is
quite close to Fisher and Simmons and is clearly more conservative at the frequencies shown.
Sehgal (2009) shows that at higher frequencies above 300kHz, Thorps model predicts lower
losses as it does not take into account the relaxation frequencies found by Fisher and Simmons.
If depth and frequency are fixed and temperature varied from 0 to 27 ◦ C, there is a decrease
in α of approximately 4 dB/km for frequencies in the range of 30 to 60kHz which correlates
to work presented by Urick (Urick, 1967)(Fig5.3 pg 89). If we consider where AUV swarms
are most likely to operate, in the ’mixed surface layer’, where temperature varies considerable
due to latitude (but has an average temp of 17◦ C (Johnson, 2011)), temperature may be an
important factor. It should be noted that if operating in lower temperatures α is higher and
thus using 0◦ C will be a conservative alternative. At shorter ranges, the significance of α is
expected to be less due to the linear relationship with range which will be discussed further
in this chapter.
As mentioned, depth (pressure) has less of an effect on α than temperature at these
lower frequencies. Domingo (Domingo, 2008) investigates the effect of depth (pressure) on
absorption and confirms that for lower frequencies of less than 100kHz there is less change in
α. More specifically Urick (1967) defined the variation by: αd = α ∗ 10−3 (1 − 5.9 ∗ 10−6 ) ∗ d
dB/m (where d = depth in meters) but has also suggested an approximation of a 2% decrease
for every 300 m depth. Thus, depth (pressure) variations are not expected to play a significant
role in short range AUV swarm operations especially those that use a 2D horizontal topology
as described in this chapter.

3.3 Path loss

Total path loss is the combined contribution of both the spreading and absorption losses.
Urick (1967) established that this formula of spreading plus absorption yields a reasonable
Short-Range Underwater Acoustic Communication Networks 1819
Short-Range Underwater Acoustic
Communication Networks

agreement with long range observations.

PathLoss(r, f , d, t) = k ∗ 10log(r ) + α( f , d, t) ∗ r ∗ 10−3 (7)

(a) Signal Attenuation showing spherical (b) Comparing Absorption Models using spherical
spreading and absorption factors spreading. Frequency change shown using Thorp
Model and Temperature ◦ ’C’ and Depth ’m’ changes
shown in Fisher and Simmons Model.

Fig. 5. Path Loss vs Range

For very short range communication (below 50 m), see Figure 5(a), the contribution of the
absorption term is less significant than the spreading term. It can be seen in Figure 5(b) that
the Thorp model shows a conservative or worst case value for the ranges of interest up to
500 m. The Fisher and Simmons model for a particular frequency however provides some
insight into the variations due to depth and temperature. However, the spreading factor k has
the most significant affect on Path Loss, seen in Figure 5(a), at these shorter ranges according
to these models.
As range increases and the absorption term begins to dominate, any variations in α also
becomes more significant. For data communication, the changes in the attenuation due to
signal frequency are particularly important as the use of higher frequencies will potentially
provide higher data rates.
In summary using the two models, Thorp and Fisher and Simmons, the two important
characteristics that can be drawn from Path Loss at the short ranges of interest for AUV swarm
operation are:
• spreading loss dominates over absorption loss, and thus the ’k’ term has a significant
impact on the attenuation of the signal at shorter ranges as illustrated in Figure 5 (a). For
AUV swarm operations and while the range between vehicles is much less than depth
spherical spreading can be assumed, and
• at the ranges below 500 m the frequency component of absorption loss is most significant
compared with the possible temperature and pressure (depth) changes as seen in Figure
5(b) and as range increases the difference also increases, effectively meaning that the
communication channel is band-limited and available bandwidth is a decreasing function
of range.
182
10 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

3.4 Underwater multipath characteristics

Multipath signals, in general, represent acoustic energy loss, however, for communication
systems it is the Inter Symbol Interference (ISI) that will also be detrimental at the receiver as
it can significantly increase the error rate of the received signal. Multipath signals are created
underwater through various mechanisms described in this section, so that, at the receiver
many components of the original signal will arrive at different times due to the different length
of propagation paths the multipath signals have taken. It is this delay spread of the signal
component arrivals that can cause ISI to occur if they overlap with previous or future signal
arrivals which will cause symbol corruption or loss and therefore bit errors. As the speed of
sound propagation is very slow in an acoustic channel this delay spread can be significant.
There are two major mechanisms responsible for creating multi-path signals and these are:
reverberation, which refers to the reflections and scattering of the sound signal; and ray
bending, which is a result of the unique sound speed structure in the oceans which create
temperature gradient channels that trap acoustic signals. Multi-path signal formation is
therefore determined by the geometry of the channel in which transmission is taking place,
the location of the transmitter and receiver, and importantly the depth at which it is occurring.
In shallow water, multi-path is due predominately to reverberation whereas in deep water it is
dominated by ray bending, although reverberation will occur in deep water if the transmitter
and receiver are located near the surface or bottom (Coates, 1989; Domingo, 2008; Etter, 2003;
Urick, 1967).
There are several physical effects which create reverberation underwater;
• Multi-path propagation caused by boundary reflections at the sea-floor or sea-surface, seen
in Figure 2.
• Multi-path propagation caused by reflection from objects suspended in the water, marine
animals or plants or bubbles in the path of the transmitted signal
• Surface scattering caused by sea-surface (waves) or sea-floor roughness or surface
absorption, particularly on the sea bottom depending on material
• Volume scattering caused by refractive off objects suspended in the signal path
Ray bending, causes various propagation path loss mechanisms in deep water depending on
the placement of transmitter and receivers. The propagating acoustic signal bends according
to Snell’s Law, to lower signal speed zones. Figure 6 shows a typical ocean Sound Speed
Profile, although variations occur with location and seasons. The profile is depth dependent,
where sound speed is influenced more by temperature in the surface layers and by pressure
at greater depths.
The various path loss mechanisms include; (Domingo, 2008)
• Surface duct, Figure 7(a) occurs when the surface layer has a positive temperature gradient,
the acoustic signals can bend back towards the surface, then reflect back into the layer off
the surface
• Deep Sound Channel, sometimes referred to as the SOFAR (Sound Fixing and Ranging)
channel, where acoustic propagation occurs above and below the level of minimum sound
speed, when the sound rays continually are bent towards the depth of minimum speed,
shown in Figure 7(a)
• Convergence zone, in deep water areas when the transmitter is located quite close to the
surface and the sound rays bend downwards as a result of decreasing temperatures until
the increase in pressure forces the rays back towards the surface, as shown in Figure 7(b)
Short-Range Underwater Acoustic Communication Networks 183
Short-Range Underwater Acoustic
Communication Networks 11

Speed of Sound
1480 m/s 1495 m/s 1510 m/s 1525 m/s
Surface Layer (Changing
0 conditions)
100
Seasonal Changes
200
Main Thermocline
(Temperature decreases rapidly)
900

Deep Isothermal Layer

1800 (Constant temperature 4°C)

Depth

2700

Fig. 6. Typical Sound Speed Profile in the Ocean Cox (1974)

• Reliable acoustic path, which occurs when the transmitter is located in very deep water
and receiver in shallow water. Referred to as reliable as it is not generally affected by
bottom or surface reflections, as shown in Figure 7(b) and
• Shadow zones that are considered a special case, as these ’zones’ are void from any signal
propagation. This means that in Shadow zones a hydrophone may not receive any signal
at all.
Thus the geometry of the channel being used is a major determinate of the number of
significant propagation paths and their relative strengths and delays. Apart from the Shadow
Zones where no signal or multipath components of the signal can reach the hydrophone, the
hydrophone may receive the direct signal and a combination of various multipath signals
that have been reflected, scattered or bent. It is these multiple components of the signal that
are delayed in time due to the various path lengths that may create ISI and errors in symbol
detection.

Ocean Surface
Ocean Surface
Surface Duct Convergence Zone

Projector Hydrophone

Deep Sound Channel

Reliable Acoustic Path

Projector Hydrophone
Deep Ocean Deep Ocean

(a) Surface Duct and Deep Sound Channel (b) Convergence Zone and Reliable Acoustic
Path

Fig. 7. Ray Bending Path Loss Mechanisms

184
12 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

For very short range channels that will be used in AUV swarm operations, multipath will
be influenced also by the range-depth ratio, which is expected to produce fewer multipath
signals at the hydrophone (Hajenko & Benson, 2010; Parrish et al., 2007). In addition some
improvement can be gained through directing the beam of the transmitted signal and the
directional properties of the receiver (Essebbar et al., 1994), however this will require an
additional level of complexity for mobile AUV’s due to the need for vehicle positioning before
sending or when receiving a signal.
Most of the discussion so far has focused on time-invariant acoustic channel multipath where
deterministic propagation path models have been developed for the various reflective and
ray bending path options. These are significant in themselves with multipath spreads in
the order of 10 to 100 ms. Take Figure 2, where projector and hydrophone are separated
by 100m and are at a depth of 100m, the delay spread between the direct path and the
first surface reflection is ≈ 28 ms. Multipath in an underwater channel, however, also
has time-varying components caused by the surface or volume scattering or by internal
waves in deep water that are responsible for random signal fluctuations. Unlike in radio
channels, the statistical characterisation of these random processes in the underwater channel
are in their early development stages. Experimental results have shown that depending
on the day, the location and the depth of communication link, the results of multipath can
follow one of the deterministic models discussed here to worst case coherence times in the
order of seconds(Stojanovic, 2006). Another source of time variability in an underwater
communication channel occurs when there is relative motion between the transmitter and
receiver as will be briefly discussed in the following sub-section.

3.5 The doppler effect

The motion of AUV’s relative to each other will cause two possible forms of Doppler distortion
in the received signal, Doppler Shifting caused by an apparent shift in frequency as the
vehicles move towards or away from each other and Doppler Spreading or its time domain
dual coherence time, which is the measure of the time varying nature of the frequency
dispersiveness in the doppler spectrum (Rappaport, 1996). The doppler shift (Δ f ) of a received
signal is f c Δv
c where f c is the original signal frequency and Δv is the relative velocity between
the moving vehicles. As an example, if the vehicles were moving at a moderately slow speed
of 1 m/s (2 knots) relative to each other and f c = 40kHz the Δ f ≈ 27Hz. Doppler spread
or coherence time measurements as mentioned above can be as long as 1 s. Thus doppler
shifting and spreading cause complications for the receiver to track the time varying changes
in the channel which need to be designed into the channel estimation algorithms and explicit
delay synchronisation approach within communication protocols. As swarm operations for
exploration require rigid topology where there is minimal relative speed differences between
vehicles, the impact of doppler effects diminished somewhat in this context and thus will not
be considered further.

3.6 Noise
There are three major contributors to noise underwater: ambient or background noise of the
ocean; self noise of the vehicle; and intermittent noise including biological noises such as
snapping shrimp, ice cracking and rain. An accurate noise model is critical to the evaluate the
SNR at the hydrophone so that the bit error rates (BER) can be establish to evaluate protocol
performance.
Short-Range Underwater Acoustic Communication Networks 185
Short-Range Underwater Acoustic
Communication Networks 13

3.6.1 Ambient noise

Ambient Noise in the ocean has been well defined (Urick, 1967). It can be represented as
Gaussian and having a continuous power spectrum density (psd). It is made up of four
components (outlined below), each having a dominating influence in different portions of
the frequency spectrum.

Fig. 8. Power Spectral density of the Ambient Noise, W - wind, S - shipping

For the frequency region of interest for AUV swarm communication systems (10 kHz to
100 kHz), the ambient noise psd decreases with increasing frequency, refer to Figure 8. At
a frequency over 100kHz the ambient thermal noise component begins to dominate and the
overall noise psd begins to increase, but this point moves further away from the frequencies
of interest for AUV communication particularly as the wind speed increases.
• Turbulence noise influences only the very low frequency regions f < 10Hz
10logNturb ( f ) = 17 − 30log( f )
• Shipping noise dominates the 10 - 100Hz region and has defined a shipping activity factor
of s whose value ranges from 0 to 1 for low to high activity respectively:
10logNship ( f ) = 40 + 20(s − 0.5) + 26log( f ) − 60log( f + 0.03)
• Wave and other surface motion caused by wind and rain is a major factor in the mid
frequency region of 100Hz - 100kHz where wind speed is given by w in m/s:
10logNwind ( f ) = 50 + 7.5w1/2 + 20log( f ) − 40log( f + 0.4)
• Thermal noise becomes dominate over 100kHz:
10logNth ( f ) = −15 + 20log( f )
where wind speed is given by w in m/s (1m/s is approximately 2 knots) and f is in kHz.
Ambient Noise power also decreases with increasing depth as the distance from the surface
and therefore shipping and wind noise becomes more distant. Ambient noise has been shown
to be 9dB higher in shallow water than deep water (Caruthers, 1977). Swarm operations,
as well as other underwater networking operations will mean that communication nodes
including AUV’s will be working in relatively close proximity to other nodes which will add
an additional level of ambient noise to their operations due to the noise of the other vehicles in
186
14 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

the swarm, irrespective of the operating depth. As will be discussed in the next section on Self
Noise, the expectation is that this additional ’ambient noise’ which relates to the ’Shipping
Noise’ component of ambient noise will have limited affect on the acoustic communication
which generally uses frequencies above 10kHz.

3.6.2 Self noise

Self noise is defined as the noise generated by the vehicle itself as the platform for receiving
signals. This noise can reach the hydrophone mounted on the AUV either through the
mechanical structure or through the water passing over the hydrophone. The degree to
which turbulent flows cause transducer self noise depends on the location (mounting) of the
transducer and its directivity characteristics (Sullivan & Taroudakis, 2008). Self Noise can
also be seen as an equivalent isotropic noise spectrum as presented by Urick from work done
during WWII on submarines. In general, as with ambient noise, there is decreasing levels
of self noise with increases in frequency however self noise is also significantly affected by
speed with decreasing noise spectra when the vessels are travelling at slower speeds or are
stationary (Eckart, 1952; Kinsler et al., 1982; Urick, 1967).
Kinsler (1982) notes that at low frequencies (<1kHz) and slow speeds machinery noise
dominates and at very slow speeds self noise is usually less important than ambient noise.
Whereas at higher frequencies (10kHz) propeller and flow noise begins to dominate and
as speed increases the hydrodynamic noise around the hydrophone increases strongly and
becomes more significant than the machinery noise. This is due to the cavitation from the
propeller due to the entrainment of air bubbles under or on the blade tip of the propeller. At
higher speeds, self noise can be much more significant than ambient noise and can become
the limiting factor.
The self noise of different size and types of vehicles are as varied as there are vehicle designs
and there is little recent published values. Each vehicle itself produces large variations in
self noise with speed and operating conditions (Eckart, 1952). Self noise can be controlled
by selection of motor type, configuration, mounting and motor drivers. The trend for most
AUV’s will be the use of small brush-less DC electric motors which have been used on the
development of the SeaVision vehicle (Mare, 2010). Preliminary testing of self noise on these
vehicles shows an increase in noise due to increases in speed, as has been predicted, but there
was no way to distinguish between machinery and hydrodynamic affects. Higher frequency
components (up to 20kHz) were present as the speed increased due to the increased work
from the thrusters. When the SeaVision vehicle hovered in a stationary position the frequency
of the noise psd centred around 2kHz, which is out of band noise.
Holmes (Holmes et al., 2005) at WHOI recently investigated the self noise of REMUS, their
torpedo shape AUV, used as a towed array. At the maximum RPM of the AUV, the 1/3rd
octave noise level, when converted to source level by the calibrated transmission factor, was
130 dB re 1μPa at 1m directly aft of the vehicle for a centre frequency of 1000 Hz. This would
represent the radiated noise source level for a vehicle moving at 3 knots (1.5m/s). Vehicles
typically radiate less noise in free operating conditions than in tethered conditions, so the
second trail on the REMUS was measuring the radiated noise of the vehicle to examine the
power spectral density of the noise as recorded on the hydrophone array as it was towed
behind the vehicle. The results showed the RPM dependent radiated noise in the aft direction
at a distance of 14.6 m behind the vehicle looking at frequency range up to 2500Hz which is
again out of band noise.
Short-Range Underwater Acoustic Communication Networks 187
Short-Range Underwater Acoustic
Communication Networks 15

As the operating frequencies of the communication system is likely to be higher than most self
noise, and the vehicles will operate relatively slowly, the expected contribution of self noise
on the hydrophone reception will be low.

3.6.3 Intermittent sources of noise

The sources of intermittent noise can become very significant in locations or times that
they occur close to operating AUV swarms. The two major areas where research has been
undertaken are in the marine bio-acoustic fields and also the effect of rain and water bubbles
created by raindrops.
Major contributors to underwater bio-acoustic noise include;
• Shellfish - Crustacea - most important here are the snapping scrimp who produce a broad
spectrum of noise between 500Hz and 20kHz
• Fish - toadfish 10 to 50Hz
• Marine mammals - Cetacea - porpoises 20 to 120Hz
Rain creates different noise spectrum to wind and needs to be dealt with separately as it is not
a constant source of noise. Urick (1967) showed examples of increases of almost 30dB in the
5 to 10kHz portion of spectrum in heavy rain, with steady rain increasing noise by 10dB or
sea state equivalent increase from 2 to 6. Eckart (1952) presented average value of rain at the
surface from 100Hz to 10kHz of -17 to 9dB.
These main contributors to intermittent sources predominate in the lower frequency ranges
up to 20kHz. Thus, interference in the operating frequencies of communication data signals is
considered low.

4. Short range channel modelling

Utilising the full capacity of the underwater acoustic channel is extremely important as the
channel exhibits such challenging and limited resources as has been discussed. For short range
data transmission operations there are a number of benefits that can be gained over current
longer range underwater acoustic transmission. These will now be explored further in terms
of data communication protocol design and development. In particular, finding the optimum
signal frequency and bandwidth at different ranges and under various channel conditions
will be evaluated on the basis of using the best Signal-to-Noise ratio (SNR) possible at the
hydrophone. Investigation into channel capacity and BER for various possible modulation
schemes will also be analysed to set up the background to the challenges for MAC and routing
protocol design.

4.1 Frequency dependent component of SNR

The narrowband Signal-to-Noise-Ratio (SNR) observed at the receiver, assuming no
multi-path or doppler losses is given by:

SL projector ( Ptx , η, DI )
SNR(r, f , d, t, w, s, Ptx ) = (8)
PathLoss(r, f , d, t) ∑ Noise( f , w, s) × B
where B is the receiver bandwidth and the Signal Level (SL), PathLoss and Noise terms have
been previously developed.
Taking the frequency dependent portion of the SNR from Equation 8, as developed by
Stojanovic (2006), is the PathLoss(r, f , d, t) ∑ Noise( f , w, s) product. Since SNR is inversely
188
16 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

proportional to PathLoss(r, f , d, t) ∑ Noise( f , w, s) factor, the PathLoss(r, f ,d,t

1
) Noise( f ,w,s)
term is
illustrated in the Figures 9 (a) for longer ranges and Figure 9 (b) for shorter ranges. The first
of these figures, shows various ranges up to 10km using Thorps absorption model (spherical
spreading) and has been presented by several authors (Chen & Mitra, 2007; Nasri et al., 2008;
Sehgal et al., 2009; Stojanovic, 2006). Figure 9(b) highlights shorter ranges of 500 m and 100 m
and also illustrates the variation between the Thorp and Fisher and Simmons absorption loss
models developed in Section 3.2.2.

(a) Longer Ranges (b) Short Ranges with channel parameter

variations

Fig. 9. Frequency Dependent component of narrowband SNR

These figures show that there is a signal frequency where the frequency dependent component
of SNR is optimised assuming that the projector parameters, including transmitter power
and projector efficiency behave uniformly over the frequency band. The black dot at the
apex of each of the three curves based on the Thorp absorption model indicate this optimum
point. The two absorption models present similar responses and optimum frequencies. There
is a minor variation in optimum frequency at 100 m where the absorption coefficient has
a significantly lower contribution. The impact that the Ambient Noise component has on
optimum signal frequency is seen most dramatically at the 100 m range when the wind state
is changed from 0 to 2 m/s, the optimal signal frequency changes from 38 kHz to 68 kHz.
From a communication perspective, if two vehicles were operating at 100 m and 38kHz and
the wind state changed from 0 m/s to 2 m/s, there is a reduction of 9dB in the frequency
dependent component of SNR. This is not an absolute reduction in SNR as the projector
parameters and in particular the transmitter power level has not been considered here. It
does however indicate the significant impact that wind and wave action can play with data
communication underwater, and in addition this reduced SNR value does not include any
increased losses associated with the increase scattering that wave action can generate. The
impact of shipping, found in the Ambient Noise term, is not included here as its effect is minor
on signal frequencies above 10kHz. In Section 3.2.2 Figure 4, temperature variations were seen
to have the most impact on the signal frequency associated with the ranges of interest. Figure
9(b) illustrates this difference in terms of the frequency dependent component of SNR. This is
further explored in Figure 10(a) in terms of the signal frequency variations over range.
Short-Range Underwater Acoustic Communication Networks 189
Short-Range Underwater Acoustic
Communication Networks 17

The impact of changes in range can be seen if the vehicles moved from 100 m to 500 m (at
wind state 0 m/s), the optimum signal frequency to maintain highest SNR decreases from
38 kHz to ≈ 28 kHz, Figure 9(b). Reduction in signal frequency implies a potential reduction
in absolute bandwidth and with that a reduction in data rate which needs to be managed. This
will be investigated further in the next sub sections. Figure 10 (a) and (b) show the optimum
signal frequency verses range up to 500 m for the various parameters; temperature and depth,
within the Thorp and Fisher and Simmons Absorption Loss models as well as the wind in the
Ambient Noise model. The optimum frequency, decreases with increasing range due to the
dominating characteristic of the absorption loss. It can be seen in Figure 10(a) that as the range
increases there is an increasing deviation between the two models and between the parameters
within the Fisher and Simmons model. There is approximately a 2.5 kHz difference between
the models themselves at 500 m and up to 6 kHz when temperature increases are included.
When wind is included, Figure 10(b), there is a dramatic change in optimum signal frequency
at very short ranges and this difference reduces substantially over the range shown. This is
due to the increasing significance of the Absorption Loss term relative to the constant Ambient
Noise term (as it is not range dependent), which reduces the affect of the Noise term and
therefore the wind parameter. In both Figure 10(a) and (b), the Fisher and Simmons model
provides higher optimum frequencies due to the more accurate inclusion of the relaxation
frequencies of boric acid and magnesium sulphate.

(a) Comparison of Absorption Loss Parameters (b) Comparison with changes in wind (from
Ambient Noise Characteristics)

Fig. 10. Optimum Frequency determined from frequency dependent component of

narrowband SNR

4.2 Channel bandwidth

Having established that at different ranges there is an optimum signal frequency that
provides a maximum SNR, assuming constant transmitter power and projector efficiency,
there is therefore an associated channel bandwidth with these conditions for different ranges.
To determine this bandwidth a heuristic of 3dB around the optimum frequency is used.
Following a similar approach to Stojanovic (2006) the bandwidth is calculated according
to the frequency range using ±3dB around the optimum signal frequency f o (r ) which
has been chosen as the centre frequency. Therefore, the f min (r ) is the frequency when
190
18 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

PathLoss(r, d, t, f o (r )) N ( f o (r )) − PathLoss(r, d, t, f )) N ( f ) ≥ 3dB holds true and similarly for

f max (r ) when PathLoss(r, d, t, f )) N ( f ) − PathLoss(r, d, t, f o (r )) N ( f o (r )) ≥ 3dB is true. The
system bandwidth B(r,d,t) is therefore determined by:

B(r, d, t) = f max (r ) − f min (r ) (9)

Thus, for a given range, there exists an optimal frequency from which a range dependent 3dB
bandwidth can be determined as illustrated in Figure 11. The changes discussed in Section
4.1, related to changes in the optimum signal frequency with changes in range and channel
conditions such as temperature, depth and wind. These variations are reflected in a similar
manner to the changes seen here in channel bandwidth and in turn will reflect in the potential
data transmission rates. Figure 11 demonstrates that both the optimal signal frequency and the
3dB channel bandwidth decrease as range increases. The impact of changing wind conditions
on channel bandwidth is significant, however as discussed wind and wave action will also
include time variant complexities and losses not included here. Temperature increases show
an increase in channel bandwidth, at ranges of interest, due to the reduction in absorption loss
as temperature increases, which means some benefits in working in the surface layers. The
discussion here highlights that the underwater acoustic channel is severely band-limited and
bandwidth efficient modulation will be essential to maximise data throughput and essentially
that major benefits can be gained when performing data transmission at shorter ranges or in
multi-hop arrangements.

Fig. 11. Range dependent 3dB Channel Bandwidth shown as dashed lines. Where Y-axis is
the frequency dependent component of the narrowband SNR

4.3 Channel capacity

Prior to evaluating the more realistic performance of the underwater data communication
channel, the maximum achievable error-free bit rate C for various ranges of interest will be
determined using the Shannon-Hartley expression, Equation 10. In these channel capacity
calculations, all the transmitted power Ptx is assumed to be transferred to the hydrophone
except for the losses associated with the deterministic Path Loss Models developed earlier.
The Shannon-Hartley expression using the Signal-to-Noise ratio, SNR(r), defined in Equation
8, is:
C = Blog2 (1 + SNR(r )) (10)
where C is the channel capacity in bps and B is the channel bandwidth in Hz
Short-Range Underwater Acoustic Communication Networks 191
Short-Range Underwater Acoustic
Communication Networks 19

Thus using the optimum signal frequency and bandwidths at 100 m and 500 m found in
the Section 4.1 and 4.2, the maximum achievable error free channel capacities against range
are shown in Figure 12. The signal frequency and channel bandwidth values for 100 m
were f o = 37kHz and B=47kHz and for 500 m were f o = 27kHz and B=33kHz. These are
significantly higher than values currently available in underwater operations(Walree, 2007),
however they provide an insight into the theoretical limits. Two different transmitter power
levels are used, 150dB re 1μPa which is approximately 10mW (Equation 1) and 140dB re
1μPa is 1mW. Looking at the values associated with the same power level in Figure 12, the
higher channel capacities are those associated with the determined optimum frequency and
bandwidth for that range as would be expected. The change in transmitter power, however,
by a factor of 10, does not produces a linear change in channel capacity across the range.
These variations are important to consider as minimising energy consumption will be critical
for AUV operations. In general, current modem specifications indicate possible data rate
capacities of less than 10kbps (LinkQuest, 2008) for modem operations under 500 m, well
short of these theoretical limits. This illustrates the incredibly severe data communication
environment found underwater and that commercial modems are generally not yet designed
to be able to adapt to specific channel conditions and varying ranges. The discussion here is
to understand the variations associated with the various channel parameters at short range
that may support adaptability and improved data transmission capacities.
Maximum Channel Capacity Achievable (kbps)

250 Ptx=1mW, Fo=27kHz

Ptx=1mW, Fo=37kHz
200 Ptx=10mW, Fo=27kHz
Ptx=10mW, Fo=37kHz

150

100 6dB 16dB

8dB 18dB

50 -3dB 7dB
-6dB 4dB SNR (dB) Values
-9dB -1dB
-13dB -2.5dB
0
100 300 500
Range (m)

Fig. 12. Theoretical limit of Channel Capacity (kbps) verse Range

4.4 BER in short range underwater acoustic communication

Achieving close to the maximum channel capacities as calculated in the previous section is
still a significant challenge in underwater acoustic communication. The underwater acoustic
channel presents significant multipaths with rapid time-variations and severe fading that
lead to complex dynamics at the hydrophone causing ISI and bit errors. The probability
of bit error, BER, therefore provides a measure of the data transmission link performance.
In underwater systems, the use of FSK (Frequency Shift Keying) and PSK (Phase Shift
Keying) have occupied researchers approaches to symbol modulation for several decades.
One approach is using the simpler low rate incoherent modulation frequency hopping FSK
192
20 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

signalling with strong error correction coding that provides some resilience to the rapidly
varying multipath. Alternatively, the use of a higher rate coherent method of QPSK signalling
that incorporates a Doppler tolerant multi-channel adaptive equalizer has gained in appeal
over that time (Johnson et al., 1999).
The BER formulae are well known for FSK and QPSK modulation techniques (Rappaport,
Eb
1996), which require the Energy per Bit to Noise psd, N o
, that can be found from the SNR
(Equation 8) by:

Eb Bc
= SNR(r ) × (11)
No Rb
where Rb is the data rate in bps and Bc is the channel bandwidth. Equation 12 and 13 are the
uncoded BER for BPSK/QPSK and FSK respectively:

1 E
QPSK : BER = er f c[ b ]1/2 (12)
2 No
1 1 Eb 1/2
FSK : BER = er f c[ ] (13)
2 2 No

Eb Eb
(a) BER vs No (b) BER vs Range vs No (for QPSK)

Fig. 13. Probability of Bit Error for Short Range Acoustic Data Transmission Underwater

The data rates Rb used are 10 and 20 kbps to reflect the current maximum commercial
Eb
achievable levels. Figure 13 (a) and (b) show the BER for No
and Range respectively. Taking a
BER of 10−4 or 1 bit error in every 10, 000 bits, the N
Eb
o
required for QPSK is 8dB for a transmitter
power of 10mW and a data rate of 20kbps. This increases to 12dB if using FSK with half
the data rate (10 kbps) and same Transmitter Power. From Figure 13 (b), these settings will
provide only a 150 m range. The range can be increased to 250 m using QPSK if the data rate
was halved to 10 kbps or out to 500 m if the transmitter power was increased to 100mW in
addition to the reduced data rate. Transmitter power plays a critical role, as illustrated here,
by the comparison of ranges achieved from ≈ 75 m to 500 m with a change of transmitter
power needed from 1mW to 100mW for this BER.
Short-Range Underwater Acoustic Communication Networks 193
Short-Range Underwater Acoustic
Communication Networks 21

5. Swarm network protocol design techniques

A short range underwater network, as shown in Figure 1(b) is essentially a multi-node sensor
network. To develop a functional sensor network it is necessary to design a number of
protocols which includes MAC, DLC (Data Link Control) and routing protocols. A typical
protocol stack of a sensor network is presented in Figure 14. The lowest layer is the physical
layer which is responsible for implementing all electrical/acoustic signal conditioning
techniques such as amplifications, signal detection, modulation and demodulation, signal
conversions, etc.. The second layer is the data link layer which accommodates the MAC and
DLC protocols. The MAC is an important component of a sensor networks protocol stack, as
it allows interference free transmission of information in a shared channel. The DLC protocol
includes the ARQ (Automatic Repeat reQuest) and flow control functionalities necessary for
error free data transmission in a non zero BER transmission environment. Design of the DLC
functionalities are very closely linked to the transmission channel conditions. The network
layers main operational control is the routing protocol; responsible for directing packets
from the source to the destination over a multi-hop network. Routing protocols keep state
information of all links to direct packets through high SNR links in order to minimise the end
to end packet delay. The transport layer is responsible for end to end error control procedures
which replicates the DLC functions but on an end to end basis rather than hop to hop basis
as implemented by the DLL. The transport layer could use standard protocols such as TCP
(Transmission Control Protocol) or UDP (User Datagram Protocol). The application layer
hosts different operational applications which either transmit or receive data using the lower
layers. To develop efficient network architectures, it is necessary to develop network and/or
application specific DLL and network layers. The following subsections will present MAC
and routing protocol design characteristics required for underwater swarm networking.

Application

Transport

Network

Data Link Layer (DLL)

Physical

Fig. 14. A typical protocol stack for a sensor network

5.1 MAC protocol

Medium access protocols are used to coordinate the transmission of information from
multiple transmitters using a shared communication channel. MAC protocols are designed
to maximise channel usage by exploiting the key properties of transmission channels. MAC
protocols can be designed to allocate transmission resources either in a fixed or in a dynamic
manner. Fixed channel allocation techniques such as Frequency Division Multiplexing (FDM)
or Time Division Multiplexing (TDM) are commonly used in many communication systems
where ample channel capacity is available to transmit information (Karl & Willig, 2006).
For low data rate and variable channel conditions, dynamic channel allocation techniques
194
22 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

are generally used to maximise the transmission channel utilisation where the physical
transmission channel condition could be highly variable. Based on the dynamic channel
allocation technique it is possible to develop two classes of MAC protocols known as random
access and scheduled access protocol. The most commonly used random access protocols is
the CSMA (Carrier Sense Multiple Access) widely used in many networks including sensor
network designs. Most commonly used scheduled access protocol is the polling protocol. Both
the CSMA and polling protocols have flexible structures which can be adopted for different
application environments. As discussed in this chapter, the underwater communication
channel is a relatively difficult transmission medium due to the variability of link quality
depending on location and applications. Also, the use of an acoustic signal as a carrier will
generate a significant delay which is a major challenge when developing a MAC protocol.
In the following subsection we discuss the basic design characteristics of the standard
CSMA/CA protocol and its applicability for underwater applications.

5.1.1 CSMA/CA protocol

Carrier Sense Multiple Access with Collision Detection protocol is a distributed control
protocol which does not require any central coordinator. The principle of this protocol is
that a transmitter that wants to initiate a transmission, checks the transmission channel by
checking the presence of a carrier signal. If no carrier signal is present which indicates the
channel is free and the transmitter can initiate a transmission. For a high propagation delay
network such a solution does not offer very high throughput due to the delay.

Node A Node B

Distance = d (m), Propagation delay = tp

Fig. 15. CSMA/CA protocol based packet transmission example

Consider Figure 15, where two nodes are using CSMA/CA protocol, are spaced apart by 100
meters. In this case, if at t=0, Node A senses the channel then it will find the channel to
be free and can go ahead with the transmission. If Node A starts transmission of a packet
immediately then it can assume that the packet will be successfully transmitted. However,
if Node B starts sensing the channel before the propagation delay time t p then it will also
find the channel is free and could start transmission. In this case both packet will collide and
the transmission channel capacity will be wasted for a period of L+t p where L is the packet
transmission time. On the other hand, if Node B checks the channel after time t p from the
commencement of A’s packet transmission, then it will find the channel is busy and will not
transmit any packets. Now this simple example shows how the performance of random access
protocol is dependent on the propagation delay. If propagation delay is small then there is
much lower probability that a packet will be transmitted before the packet from A arrives at
B. As the propagation delay increases the collision probability will also increase.
The CSMA/CA protocol is generally used in RF (Radio Frequency) networks where 100 m
link delay will incur a propagation delay of 0.333 μsec whereas an underwater acoustic link
Short-Range Underwater Acoustic Communication Networks 195
Short-Range Underwater Acoustic
Communication Networks 23

of same distance will generate a propagation delay of 0.29 sec which is about 875,000 times
longer than the RF delay. One can easily see why an acoustic link will produce much lower
throughput than is predicted by the Shannon-Hartley theorem as discussed in Section 4.3. If
we assume that we are transmitting a 100 byte packet, then the packet will take about 0.08 sec
to transmit on a 10 kbps RF link. The same packet will take 0.3713 sec on a 10 kbps acoustic
link offering a net throughput of 2.154 kbps. This calculation is based on the assumption that
the transmission channel is ideal i.e. BER=0. If the BER of the channel is non zero then the
throughput will be further reduced.
Previous sections have shown that the BER of a transmission link is dependent on the link
parameters, geometry of the application environment, modulation techniques, and presence
of various noise sources. Non zero BER conditions introduce a finite packet error rate (PER)
on a link which is described by Equation 14, where K represents the packet length. The PER
will depend on the BER and the length of the transmitted packet. For a BER of 10−3 using a
packet size of 100 bytes, the link will generate a PER value of 0.55 which means that almost
every second packet will be corrupted and require some sort of error protection scheme to
reduce the effective packet error rate.
There are generally two types of packet error correction techniques used in communication
systems, one is forward error correction (FEC) scheme which uses a number of redundant bits
added with information bits to offer some degree of protection against the channel error. The
second technique involves the use of packet retransmission techniques using the DLC function
known as the ARQ. The ARQ protocol will introduce retransmissions when a receiver is
unable to correct a packet using the FEC bits. The retransmission procedure could effectively
reduce the throughput of a link further because the same information is transmitted multiple
times. From this brief discussion one can see that standard CSMA/CA protocols used in
sensor networks are almost unworkable in the underwater networking environment unless
the standard protocol is further enhanced. This is a major research issue which is currently
followed up by many researchers and authors. Readers can find some of the current research
work on the MAC protocol in the following references (Chirdchoo et al., 2008; Guo et al., 2009;
Pompili & Akyildiz, 2009; Syed et al., 2008).

PER = 1 − (1 − BER)K (14)

5.2 Packet routing

Packet routing is another challenging task in the underwater networking environment. Packet
routing protocols are very important for a multi-hop network because the receivers and the
transmitters are distributed in a geographical area where nodes can also change their positions
over time. Each node maintains a routing table to forward packets through multi-hop links.
Routing tables are created by selecting the best cost paths from transmitters to receivers. The
cost of a path can be expressed in terms of delay, packet loss, BER, real monetary cost $,
etc.. For underwater networks, the link delay could be used as a cost metric, to transmit
packets with a minimum delay. Routing protocols are generally classified into two classes:
distance vector and link state routing protocols (LeonGarcia & Widjaja, 2004). The distance
vector algorithms generally select a path from a transmitter to receiver based on shortest
path through neighbouring networks. When the status of a link changes, for example, if the
delay or SNR of a link is increased then the node next to the link will detect and inform its
neighbour about the change and suggest a new link. This process will continue until all the
nodes in the network have updated their routing table. The link state routing protocols work
196
24 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

in a different manner. In this case all the link state information is periodically transmitted to
all nodes in the network. In case of any change of state of a link, all nodes get notification
and modify their routing table. In a swarm network link qualities will be variable which will
require regular reconfiguration of routing tables. The performance of routing algorithms is
generally determined by a number of factors including the convergence delay. In the case of
a swarm network the convergence delay will be a critical factor because of high link delays.
For underwater swarm applications, each update within a network will take considerably
longer time than a RF network, causing additional packet transmission delays. Hence, it
is necessary to develop the network structure in different ways than a conventional sensor
network. For example, it may be necessary to develop smaller size clustered networks where
cluster heads form a second tier network. Within this topology, local information will flow
within the cluster and inter-cluster information will flow through the cluster head network.
Cluster based communication architectures are also being used in Zigbee based and wireless
personal communication networks (Karl & Willig, 2006). Further research is necessary to
develop appropriate routing algorithms to minimise packet transmission delay in swarm
networks. Readers can consult the following references to follow some of the recent progress
in the area (Aldawibio, 2008; Guangzhong & Zhibin, 2010; LeonGarcia & Widjaja, 2004; Zorzi
et al., 2008).
Discussion in this section clearly shows that the MAC and routing protocol designs require
transmission channel state information in order to optimise their performance. Due to the
high propagation delay of an underwater channel, any change of link quality such as SNR
will significantly affect the performance of the network. Hence, it is necessary to develop a
new class of protocols which can adapt themselves with the varying channel conditions and
offer reasonable high throughput in swarm networks.

6. Conclusion
The increasing potential of Autonomous Underwater Vehicle (AUV) swarm operations and
the opportunity to use multi-hop networking underwater has led to a growing need to
work with a short-range acoustic communication channel. Understanding the channel
characteristics for data transmission is essential for the development and evaluation of new
MAC and Routing Level protocols that can better utilise the limited resources within this
harsh and unpredictable channel.
The constraints imposed on the performance of a communication system when using
an acoustic channel are the high latency due to the slow speed of the acoustic signal
(compared with RF), and the signal fading properties due to absorption and multipath
signals, particularly due to reflections off the surface, sea floor and objects in the signal path.
The shorter range acoustic channel has been shown here to be able to take advantage of
comparatively lower latency and transmitter power as well as higher received SNR and signal
frequencies and bandwidths (albeit still only in kHz range). Each of these factors influence the
approach needed for developing appropriate protocol designs and error control techniques
while maintaining the required network throughput and autonomous operation of each of
the nodes in the swarm.
Significant benefits will be seen when AUVs can operate as an intelligent swarm of
collaborating nodes and this will only occur when they are able to communicate quickly and
clearly between each other in a underwater short range ad-hoc mobile sensor network.
Short-Range Underwater Acoustic Communication Networks 197
Short-Range Underwater Acoustic
Communication Networks 25

7. References
Aldawibio, O. (2008). A review of current routing protocols for ad hoc underwater acoustic
networks, First International Conference on the Applications of Digital Information and
Web Technologies. ICADIWT, pp. 431 –434.
Caruthers, J. (1977). Fundamentals of Marine Acoustics, Elsevier Scientific Publishing.
Chen, W. & Mitra, U. (2007). Packet scheduling for multihopped underwater acoustic
communication networks, IEEE OCEANS’07, pp. 1–6.
Chirdchoo, N., Soh, W. & Chua, K. (2008). Ript: A receiver-initiated reservation-based protocol
for underwater acoustic networks, IEEE Journal on Selected Areas in Communications
26(9): 1744 –1753.
Coates, R. (1989). Underwater Acoustic Systems, John Wiley and Sons.
Cox, A. W. (1974). Sonar and Underwater Sound, Lexington Books.
Domingo, M. (2008). Overview of channel models for underwater wireless communication
networks, Physical Communication pp. 163 – 182.
Dunbabin, M., Roberts, J., Usher, K., Winstanley, G. & Corke, P. (2005). A hybrid auv
design for shallow water reef navigation, Proc. International Conference on Robotics and
Automation (ICRA), pp. 2117–2122.
Eckart, C. (1952). Principles of Underwater Sound, Research Analysis Group, National Research
Council, California University.
Essebbar, A., Loubet, G. & Vial, F. (1994). Underwater acoustic channel simulations for
communication, IEEE OCEANS ’94. ’Oceans Engineering for Today’s Technology and
Tomorrow’s Preservation.’, Vol. 3, pp. III/495 –III/500 vol.3.
Etter, P. (2003). Underwater Acoustic Modeling and SImulation, third edn, Spon Press.
Fisher, F. & Simmons, V. (1977). Sound absorption in sea water, Journal of the Acoustical Society
of America 62(3).
Francois, R. & Garrison, G. (1982). Sound absorption based on ocean measurements: Part 1
and 2, Journal of the Acoustical Society of America 72(3,6): 896–907, 1879 – 1890.
Guangzhong, L. & Zhibin, L. (2010). Depth-based multi-hop routing protocol for underwater
sensor network, 2nd International Conference on Industrial Mechatronics and Automation
(ICIMA), Vol. 2, pp. 268 –270.
Guo, X., Frater, M. & Ryan, M. (2009). Design of a propagation-delay-tolerant mac protocol for
underwater acoustic sensor networks, IEEE Journal of Oceanic Engineering 34(2): 170
–180.
Hajenko, T. & Benson, C. (2010). The high frequency underwater acoustic channel, IEEE
OCEANS 2010, Sydney, pp. 1 –3.
Holmes, J., Carey, W., Lynch, J., Newhall, A. & Kukulya, A. (2005). An autonomous
underwater vehicle towed array for ocean acoustic measurements and inversions,
IEEE Oceans 2005 - Europe, Vol. 2, pp. 1058 – 1061 Vol. 2.
Johnson, M., Preisig, J., Freitag, L.& Stojanovic, M. (1999). FSK and PSK performance of the
utility acoustic modem, IEEE OCEANS ’99 MTS. Riding the Crest into the 21st Century,
Vol. 3, pp. 1512 Vol. 3.
Johnson, R. (2011). University Corporation of Atmospheric Research, Window to the Universe,
University Corporation of Atmospheric Research, http://www.windows. ucar. edu/
tour/link/earth/Water/overview.html.
Karl, H. & Willig, A. (2006). Protocols and Architecures for Wireless Sensor Networks, John Wiley
and Sons, Ltd.
198
26 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

Kinsler, L., Frey, A., Coppens, A. & Sanders, J. (1982). Fundementals of Acoustics, John Wiley
and Sons.
LeonGarcia, A. & Widjaja, I. (2004). Communication Networks: Fundamental Concepts and Key
Architecture, second edn, McGraw Hill.
LinkQuest (2008). SoundLink Underwater Acoustic Modems, High Speed, Power Efficient, Highly
Robust, LinkQuest Inc., http://www.link-quest.com/.
Mare, J. (2010). Design considerations for wireless underwater communication transceiver,
OCEANS10, Sydney.
Nasri, N., Kachouri, A., Andrieux, L. & Samet, M. (2008). Design considerations for wireless
underwater communication transceiver, International Conference on Signals, Circuits
and Systems.
Parrish, N., Roy, S., Fox, W. & Arabshahi, P. (2007). Rate-range for an fh-fsk acoustic modem,
Proceedings of the second workshop on Underwater networks, WuWNet ’07, pp. 93–96.
Pompili, D. & Akyildiz, I. (2009). Overview of networking protocols for underwater wireless
communications, IEEE Communications Magazine 47(1): 97 –102.
Rappaport, T. (1996). Wireless Communications, Principles and Practice, Prentice Hall.
Sehgal, A., Tumar, I. & Schonwalder, J. (2009). Variability of available capacity due to the
effects of depth and temperature in the underwater acoustic communication channel,
IEEE OCEANS 2009, EUROPE, pp. 1–6.
Stojanovic, M. (2006). On the relationship between capcity and distance in an underwater
acoustic communication channel, International Workshop on Underwater Networks,
WUWNet’06.
Stojanovic, M. (2008). Underwater acoustic communications: Design considerations on the
physical layer, 5th Annual Conference on Wireless on Demand Network Systems and
Services, WONS, pp. 1–10.
Sullivan, E. & Taroudakis, M. (2008). Handbook of Signal Processing in Acoustics Volume 2,
RSpringer.
Syed, A., Ye, W. & Heidemann, J. (2008). Comparison and evaluation of the t-lohi
mac for underwater acoustic sensor networks, IEEE Journal on Selected Areas in
Communications 26(9): 1731 –1743.
Thorp, W. H. (1965). Deep-ocean sound attenuation in the sub- and low-kilocycle-per-second
region, Journal of the Acoustical Society of America 38(4): 648–654.
Urick, R. (1967). Principles of Underwater Sound for Engineers, McGraw-Hill.
Waite, A. (2005). Sonar for Practicing Engineers, third edn, Wiley.
Walree, P. (2007). Acoustic modems: Product survey, Hydro International Magazine 11(6): 36–39.
Zorzi, M., Casari, P., Baldo, N. & Harris, A. (2008). Energy-efficient routing schemes for
underwater acoustic networks, IEEE Journal on Selected Areas in Communications
26(9): 1754 –1766.
 9

Embedded Knowledge and Autonomous

Planning: The Path Towards Permanent
Presence of Underwater Networks
Pedro Patrón, Emilio Miguelañez and Yvan R. Petillot
Ocean Systems Laboratory, Heriot-Watt University
United Kingdom

1. Introduction
Oceanographic observatories, year-round energy industry subsea field inspections and
continuous homeland security coast patrolling now all require the routine and permanent
presence of underwater sensing tools.
These applications require underwater networks of fixed sensors that collaborate with
fleets of unmanned underwater vehicles (UUVs). Technological challenges related to the
underwater domain, such as power source limitations, communication and perception noise,
navigation uncertainties and lack of user delegation, are limiting their current development
and establishment. In order to overcome these problems, more evolved embedded tools
are needed that can raise the platform’s autonomy levels while maintaining the trust of the
operator.
Embedded decision making agents that contain reasoning and planning algorithms can
optimize the long term management of heterogeneous assets and provide fast dynamic
response to events by autonomously coupling global mission requirements and resource
capabilities in real time. The problem, however, is that, at present, applications are
mono-domain: Mission targets are simply mono-platform, and missions are generally static
procedural list of commands described a-priori by the operator. All this, leaves the platforms
in isolation and limits the potential of multiple coordinated actions between adaptive
collaborative agents.
In a standard mission flow, operators describe the mission to each specific platform, data is
collected during mission and then post-processed off-line. Consequently, the main use for
underwater platforms is to gather information from sensor data on missions that are static
and incapable to cope with the long term environmental challenges or resource changes.
In order for embedded service agents to make decisions and interoperate, it is necessary that
they have the capability of dealing with and understanding the highly dynamic and complex
environments where these networks are going to operate. These decision making tools are
constrained to the quality and scope of the available information.
Shared knowledge representation between embedded service-oriented agents is therefore
necessary to provide them with the required common situation awareness. Two sources
can provide this type of information: the domain knowledge extracted from the expert
(orientation) and the inferred knowledge from the processed sensor data (observation). In
both cases, it will be necessary for the information to be stored, accessed and shared efficiently
200
2 Autonomous Underwater Vehicles
Underwater Vehicles

by the deliberative agents while performing a mission. These agents, providing different
capabilities and working in collaboration, might even be distributed among the different
platforms or sharing some limited resources.

1.1 Contribution
In this chapter, we first provide a review to the different approaches solving the decision
making process for UUV missions. Then, we propose a semantic framework that provides
a solution for hierarchical distributed representation of knowledge for multidisciplinary
agent interaction. This framework uses a pool of hierarchical ontologies for representation
of the knowledge extracted from the expert and the processed sensor data. It provides a
common machine understanding between embedded agents that is generic and extendable.
It also includes a reasoning interface for inferring new knowledge from the observed
data and guarantee knowledge stability by checking for inconsistencies. This framework
improves local (machine level) and global (system level) situation awareness at all levels of
service capabilities, from adaptive mission planning and autonomous target recognition to
deliberative collision avoidance and escape. It acts as as an enabler for on-board decision
making.
Based on their capabilities, service-oriented agents can then gain access to the different
levels of information and contribute to the enrichment of the knowledge. If the required
information is unavailable, the framework provides the facility to request other agents with
the necessary capabilities to generate the required information, i.e. an target classification
algorithm could query the correspondent agent to provide the required object detection
analysis before proceeding with its classification task.
Secondly, we present an algorithm for autonomous mission adaptation. Using the knowledge
made available by the semantic framework, our approach releases the operator from decision
making tasks. We show how adaptation plays an important role in providing long term
autonomy as it allows the platforms to react to events from the environment while at the
same time requires less communication with the operator. The aim is to be effective and
efficient as a plan costs time to prepare. Once the initial time has been invested preparing the
initial plan, when changes occur, it might be more efficient to try to reuse previous efforts by
repairing it. Also, commitments might have been made to the current plan: trajectory reported
to other intelligent agents, assignment of mission plan sections to executors or assignment of
resources, etc. Adapting an existing plan ensures that as few commitments as possible are
invalidated. Using plan proximity metrics, we prove how similar plans are more likely to be
accepted and trusted by the operator than one that is potentially completely different.
Finally, we show during a series of in-water trials how these two elements combined,
a decision making algorithm and shared knowledge representation, provide the required
interoperability between embedded service-oriented agents to achieve high-level mission
goals, detach the operator from the routinary mission decision making and, ultimately, enable
the permanent presence of dynamic sensing networks underwater.

2. Unmanned decision making loop

In this section, we describe the decision making process currently used by UUV systems and
we introduce the unmanned decision loop, where observations, orientations, decisions and
actions (OODA) occur in a loop enabling adaptive mission planning.
In order to describe the unmanned decision loop, we need to start by modelling the mission
environment. A mission environment is defined by the tuple Π = (Σ, Ω), where:
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 2013

• Σ is the mission domain model containing information about domain, i.e. the platform
and the environment of execution, and
• Ω is the mission problem model containing information about the problem, i.e. mission
status, requirements, and objectives.
The set of all possible mission environments for a given domain is defined as the domain space
(e.g., the domain space of the underwater domain). It is denoted by Θ. A mission environment
Π is an element of one and only one Θ.
From this model, a mission plan π that tries to accomplish the mission objectives can be
produced. However, this mission environment evolves over time t as new observations of
the domain model Σt and the problem model Ωt continuously modify it:

Π t ← Π t −1 ∪ Σ t ∪ Ω t (1)
The decision making process to calculate a mission plan πt for a given mission environment
Πt occurs in a cycle of observe-orient-decide-act. This process was termed by Boyd (1992) as
the OODA-loop, and it was modelled on human behaviour. Inside this loop, the Orientation
phase contains the previously acquired knowledge and initial understanding of the situation
of the mission environment (Πt−1 ). The Observation phase corresponds to new perceptions of
the mission domain model (Σt ) and the mission problem model (Ωt ) that modify the mission
environment. The Decision component represents the level of comprehension and projection,
the central mechanism enabling adaptation before closing the loop with the Action stage.
Note that it is possible to make decisions by looking only at orientation inputs without making
any use of observations. In this case, Eq. 1 becomes Πt ← Πt−1 . In the same way, it is also
possible to make decisions by looking only at the observation inputs without making use of
available prior knowledge. In this case, Eq. 1 becomes Πt ← Σt ∪ Ωt .
In current UUVs implementations, the human operator constitutes the decision phase.
See Figure 1 for a schematic representation of the control loop. When high bandwidth
communication links exist, the operator remains in the OODA-loop during the mission
execution taking the decisions. For each update of the mission environment Πt received,
the operator decides on the correspondent mission plan πt to be performed. From the list of
actions in this mission plan, the mission executive issues the correspondent commands to the
platform. Examples of the implementation of this architecture are existing Remotely Operated
Vehicles (ROVs).
However, when communication is unreliable or unavailable, the operator must attempt to
include all possible if-then-else cases to cope with execution alternatives before the mission
starts. This is the case of current UUVs implementations that follow an orientation-only
model. Figure 2 shows this model, where the OODA-loop is broken because observations
are not reported to the human operator.
We will now discuss a few recent UUV implementations which show where the
state-of-the-art is currently positioned. Most implementations rely on pre-scripted mission
πt Commands
Π0 Human Mission
Platform Environment
Operator Executive
Events
Πt Observations

Fig. 1. Observation, Orientation, Decision and Action (OODA) loop for unmanned vehicle
systems with decision making provided by the human operator.
202
4 Autonomous Underwater Vehicles
Underwater Vehicles

π0 Commands
Π0 Human Mission
Platform Environment
Operator Executive
Events
Πt Observations

Offline Online

Fig. 2. Broken OODA-loop. Decision stage on the human operator based only on initial
pre-mission orientation.

plan managers that are procedural and static and might not even consider conditional
executions (Hagen, 2001). At this level, the mission executive follows a sequence of
basic command primitives and issues them to the functional control layer of the platform.
Description about how these approaches maintain control of underwater vehicles can be
found in Fossen (1994), Ridao et al. (1999) and Yuh (2000). In this situation, decisions taken by
the operator are made using only orientation inputs related to some previous experience and
a-priori knowledge. This has unpredictable consequences, in which unexpected situations can
cause the mission to abort and might even cause the loss of the vehicle (Griffiths, 2005; von
Alt, 2010).
More modern approaches are able to mitigate this lack of adaptability by introducing sets
of behaviours that are activated based on observations (Arkin, 1998). Behaviours divide
the control system into a parallel set of competence-levels. They can be seen as manually
scripted plans generated a-priori to encapsulate the decision loop for an individual task. Under
this approach, the key factor is to find the right method for coordinating these competing
behaviours.
The subsumption model, attributed to Brooks (1986), arbitrates behaviour priorities through
the use of inhibition (one signal inhibits another) and suppression (one signal replaces other)
networks. Most recent UUV control systems are a variant of the subsumption architecture.
This model was first applied to the control of UUVs by Turner (1995) during the development
of the ORCA system. This system used a set of schemas in a case-based framework. However,
its scalability remains unclear as trials for its validation were not conducted.
Later, Oliveira et al. (1998) developed and deployed the CORAL system based on Petri nets.
The system was in charge of activating the vehicle primitives needed to carry out the mission.
These primitives were chained by preconditions and effects.
The scaling problem was addressed by Bennet & Leonard (2000) using a layered control
architecture. Layered control is a variant of the subsumption model that restricts of interaction
between layers in order to keep it simple (Bellingham et al., 1990). The system was deployed
for the application of adaptive feature mapping.
Another approach for coordinating behaviours is vector summation that averages the action
between multiple behaviours. Following this principle, the DAMN system developed
by Rosenblatt et al. (2002) used a voting-based coordination mechanism for arbitration
implementing utility fusion with fuzzy logic.
The MOOS architecture developed by Newman (2002) was also able to guide UUVs by using a
mission control system called Helm. Helm’s mission plan was described by a set of prioritised
primitive tasks. The most suitable action was selected using a set of prioritised mission goals.
It used a state-machine for execution, a simplified version of a Petri net.
The O2 CA2 system (Carreras et al., 2007) also used a Petri net representation of the mission
plan (Palomeras et al., 2009). The system maintains the low level control (dynamics) from the
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 2035

guidance control (kinematics) uncoupled (Caccia & Veruggio, 2000). Although it contained a
declarative mission representation, missions were programmed manually.
A detailed survey of other behaviour-based approaches applied to mission control systems
for UUVs can be found in Carreras et al. (2006).
More recently, Benjamin et al. (2009) has applied multiple objective decision theory to
provide a suitable framework for formulating behaviour-based controllers that generate
Pareto-optimal and satisfying behaviours. This approach was motivated by the infeasibility of
optimal behaviour selection for real-world applications. This approach has been implemented
and deployed as part of the IvP Helm extension to MOOS. This method seems to be a
more suitable for behaviour selection, although more computationally expensive. Also, the
approach is limited to the control of only the direction and velocity parameters of the host
platform.
After reviewing this related work, two problems affecting the effectiveness of the decision
loop become evident. Firstly, orientation and observation should be linked together because
it is desirable to place the new observations in context. Secondly, decision and action should
be iterating continuously. These two problems have not been addressed together by previous
approaches. These are the two of the goals that we address in this chapter. In order to achieve
them autonomously, two additional components are required: a status monitor and a mission
plan adapter. The status monitor reports any changes detected in the mission environment
during the execution of a mission. When the mission executive is unable to handle the
changes detected by the status monitor, the mission planner is called to generate a new
modified mission plan that agrees with the updated mission environment. Figure 3 shows
the OODA-loop for autonomous decision making. Comparing it to the previous Figure 2, the
addition of status monitor and mission planner removes the need for human decisions in the
loop. Note that the original mission plan π0 could also be autonomously generated as long as
the high-level goals are provided by the human operator in Π0 .
Πt
Mission Status
Adapter Monitor

πt Observations
π0 Commands
Π0 Mission Mission
Platform Environment
Generator Executive
Events

Offline Online

Fig. 3. Required OODA-loop for autonomous decision making in UUVs. Decision stage for
adaptation takes place on-board based on initial orientation provided by the operator and
observations provided by the status monitor.
Adaptive mission planning enables a true unmanned OODA-loop. This autonomous decision
making loop copes with condition changes in the mission environment during the mission
execution. As a consequence, it releases the operator from decision making tasks in stressful
environments containing high levels of uncertainty and dynamism.
The potential benefits of adaptive mission planning capabilities for autonomous decision
making in UUVs were promoted by Turner (2005), Bellingham et al. (2006) and Patrón
& Petillot (2008). Possibly the most advanced autonomous decision making framework
for UUVs has been developed at the Monterey Bay Aquarium Research Institute. This
architecture, known as T-REX, has been deployed successfully inside the Dorado AUV (Rajan
204
6 Autonomous Underwater Vehicles
Underwater Vehicles

et al., 2009). This is now providing adaptive planning capabilities to oceanographers for
maximising the science return of their UUV missions (McGann et al., 2007; Rajan et al., 2007).
Using deliberative reactors for the concurrent integration of execution and planning (McGann,
Py, Rajan & Henthorn, 2008), live sensor data can be analysed during mission to adapt the
control of the platform in order to measure dynamic and episodic phenomenon, such as
chemical plumes (McGann, Py, Rajan, Henthorn & McEwen, 2008; McGann et al., 2009).
Alternative approaches to adaptive plume tracing can also be found in the works of Farrell
et al. (2005) and Jakuba (2007). Their research goals of all these approaches have been
motivated by scientific applications and do not consider the needs of the human operators
or the maritime industry.
However, autonomy cannot be achieved without humans, as it is necessary for this autonomy
to be ultimately accepted by an operator. Our research is geared towards improving
human access to UUVs in order to solve the maritime industry’s primary requirement of
improving platform operability (Patrón et al., 2007). We propose a goal-based approach
to solving adaptive mission planning. The advantage of this approach is that it provides
high levels of mission abstraction. This makes the human interface simple, powerful and
platform independent, which greatly eases the operator’s task of designing and deploying
missions (Patrón, 2009). Ultimately, operators will not need any specialist training for an
UUV specific platform, and instead missions will be described purely in terms of their goals.
Apart from ease of use, we have also demonstrated using a novel metric (Patrón & Birch,
2009) that adaptive mission planners can produce solutions which are close to what a human
planner would produce (Patrón et al., 2009a). This means that our solutions can be trusted by
an operator.
Another advantage of our research over other state-of-the-art UUV implementations,
is that we are industry focussed. Our service-oriented approach provides goal-based
mission planning with discoverable capabilities, which meets industry’s need for platform
independence (Patrón et al., 2009b). Finally, our plan repair approach optimises the resources
required for adaptability and maximises consistency with the original plan, which improves
human acceptance of autonomy. Resource optimisation and consistency are very important
properties for real world implementations, as we demonstrate in our sea trials (Patrón,
Miguelanez, Petillot & Lane, 2008).
Section 3 describes how do we link together orientation and observation. Section 4 presents
an approach to the continuous iteration of decision and action.

3. Semantic knowledge-based situation awareness

Unmanned vehicle situation awareness SAV consists in enabling the vehicle to autonomously
understand the ‘big picture’ (Adams, 2007). This picture is composed of the experience gained
from previous missions (orientation) and the information obtained from the sensors while
on mission (observation). Ontologies allow the representation of knowledge of these two
components.
Ontologies are models of entities and interactions, either generically or in some particular
practice of knowledge (Gruber, 1995). The main components of an ontology are concepts and
axioms. A concept represents a set or class of entities within a domain (e.g., a fault is a concept
within the domain of diagnostics). Axioms are used to constrain the range and domain of the
concepts (e.g., a driver is a software that has a hardware). The finite set of concept and axiom
definitions is called the Terminology Box TBox of the ontology.
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 2057

Fig. 4. Knowledge Base representation system including the TBox, ABox, the description
language and the reasoning components. Its interface is made of orientation rules and agent
queries.

Instances are the individual entities represented by a concept of the ontology (e.g. a
remus is an instance of the concept UUV). Relations are used to describe the interactions
between individuals (e.g. the relation isComponentOf might link the individual SensorX to
the individual PlatformY). This finite set of instances and relations about individuals is called
the Assertion Box ABox. The combination of TBox and ABox is what is known as a Knowledge
Base. TBox aligns naturally to the orientation component of SAV while ABox aligns to the
observation component.
In the past, authors such as Matheus et al. (2003) and Kokar et al. (2009) have used ontologies
for situation awareness in order to assist humans during information fusion and situation
analysis processes. Our work extends these previous works by using ontologies for providing
unmanned situation awareness in order to assist autonomous decision making algorithms
in underwater vehicles. One of the main advantages of using a knowledge base over a
classical data base schema to represent SAV is the extended querying that it provides, even
across heterogeneous data systems. The meta-knowledge within an ontology can assist an
intelligent agent (e.g., status monitor, mission planner, etc.) with processing a query. Part of
this intelligent processing is due to the capability of reasoning. This enables the publication
of machine understandable meta-data, opening opportunities for automated information
processing and analysis.
For instance, a status monitor agent using meta-data about sensor location could
automatically infer the location of an event based on observations from nearby
sensors (Miguelanez et al., 2008). Inferences over the ontology are made by reasoners. A
reasoner enables the domain’s logic to be specified with respect to the context model and
applied to the corresponding knowledge i.e., the instances of the model (see Fig. 4). A
detailed description of how a reasoner works is outside of the scope of this article. For the
implementation of our approach, we use the open source reasoner called Pellet (Sirin et al.,
2007).

3.1 Semantic knowledge-based framework

A library of knowledge bases comprise the overall knowledge framework used in our
approach for building SAV (Miguelanez et al., 2010; Patrón, Miguelanez, Cartwright &
Petillot, 2008). Reasoning capabilities allow concept consistency providing reassurance that
SAV remains stable through the evolution of the mission. Also, inference of concepts and
relationships allows new knowledge to be extracted or derived from the observed data. In
order to provide with a design that supported maximum reusability (Gruber, 1995; van Heijst
et al., 1996), we adopt a three-level segmentation structure that includes the (1) Foundation,
(2) Core and (3) Application ontology levels (see Fig. 5).
206
8 Autonomous Underwater Vehicles
Underwater Vehicles

Fig. 5. Levels of generality of the library of knowledge bases for SAV . They include the
Foundation Ontology, the Core Ontology, and the Application Ontology levels.

Foundational Ontologies (FOs) represents the very basic principles and includes Upper and
Utility Ontologies. Upper ontologies describe generic concepts (e.g., the Suggested Upper
Merged Ontology or SUMO (Niles & Pease, 2001)) while Utility ontologies describe support
concepts or properties (e.g. OGC_GML for describing geospatial information (Portele, 2007)).
FOs meet the requirement that a model should have as much generality as possible, to ensure
reusability across different domains.
The Core Ontology provides a global and extensible model into which data originating from
distinct sources can be mapped and integrated. This layer provides a single knowledge
base for cross-domain agents and services (e.g., vehicle resource / capabilities discovery,
vehicle physical breakdown, and vehicle status). A single model avoids the inevitable
combinatorial explosion and application complexities that results from pair-wise mappings
between individual metadata formats and ontologies.
In the bottom layer, an Application Ontology provides an underlying formal model for agents
that integrate source data and perform a variety of extended functions. As such, higher levels
of complexity are tolerable and the design is motivated more by completeness and logical
correctness than human comprehension. Target areas of these Application Ontologies are
found in the status monitoring of the vehicle and its environment and the planning of the
mission.
Figure 6 represents the relationship between the Foundation Ontologies (Upper and Utility),
the Core Ontology and the Application Ontology for each service-oriented agent. Raw data
gets parsed from sensors into assertions during the mission using a series of adapter modules
for each of the sensing capabilities. It also shows that the knowledge handling by the agent
during its decision making process is helped by the reasoner and the rule engine process.

Fig. 6. SAV representation in the Knowledge Base using Core and Application ontologies
supported by Upper and Utility ontologies. Generation of instances from raw data is
performed by the Adapter. Handling of knowledge is done by the Reasoner, Rule Engine and
the Service-Oriented Agent.

3.2 Foundation and core ontology

To lay the foundation for the knowledge representation of unmanned vehicles, consideration
was placed on the Joint Architecture for Unmanned Systems (JAUS) (SAE, 2008a). This
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 2079

standard was originally developed for the Unmanned Ground Vehicles (UGVs) environment
only but has recently been extended to all other environments, such as air and water, trying
to provide a common set of architecture elements and concepts. The JAUS model classifies
four different sets of Knowledge Stores: Status, World map, Library and Log. Our experience
has shown that overlap exists between these different sets of knowledge stores. The approach
proposed in this paper provides more flexibility in the way the information can be accessed
and stored, while still providing JAUS ’Message Interoperability’ (SAE, 2008b) between
agents.
Within the proposed framework, JAUS concepts are considered as the Foundation Ontology
for the knowledge representation. The Core Ontology developed in this work extends
these concepts while remaining focused in the domain of unmanned systems. Some of the
knowledge concepts identified related with this domain are:
• Platform: Static or mobile (ground, air, underwater vehicles),
• Payload: Hardware with particular properties, sensors or modules,
• Agent: Software with specific capabilities,
• Sensor: A device that receives and responds to a signal or stimulus,
• Driver: Module for interaction with a specific sensor / actuator,
Additionally, the Standard Ontology for Ubiquitous and Pervasive Applications
(SOUPA) (Chen et al., 2004) is used as an Utility Ontology. By providing generic
context-aware concepts, it enables the spatio-temporal representation of concepts in the
Core Ontology.

3.3 Application Ontology

Each service-oriented agent has its own Application Ontology. It represents the agent’s
awareness of the situation by including concepts that are specific to the expertise of the agent.
In the case study presented in this chapter, these agents are the status monitor and the mission
planner. Together, they provide the status monitor and mission adapter components described in
Fig. 3 required for closing the OODA-loop and provide on-board decision making adaptation.

3.3.1 Status Monitoring Application Ontology

The Status Monitoring Application Ontology is used to express the SAV of the status monitor
agent. To model the behaviour of all components and subsystems considering from sensor
data to possible model outputs, the Status Monitoring Application Ontology is designed and
built based on ontology design patterns (Blomqvist & Sandkuhl, 2005). Ontology patterns
facilitate the construction of the ontology and promote re-use and consistency if it is applied to
different environments. In this work, the representation of the monitoring concepts are based
on a system observation design pattern. Some of the most important concepts identified for
status monitoring are:
• Data: all internal and external variables (gain levels, water current speed),
• Observation: patterns of data (sequences, outliers, residuals,...),
• Symptom: individuals related to interesting patterns of observations (e.g., low gain levels,
high average speed),
208
10 Autonomous Underwater Vehicles
Underwater Vehicles

• Event: represents a series of correlated symptoms (low power consumption, position

drift), Two subclasses of Events are defined: CriticalEvent for high priority events and
IncipientEvent for the remaining ones.
• Status: links the latest and most updated event information to the systems being monitored
(e.g. sidescan transducer),
Please note how some of these concepts are related to concepts of the Core Ontology
(e.g. an observation comes from a sensor). These Core Ontology elements are the enablers
for the knowledge exchange between service-oriented agents. This will be shown in the
demonstration scenario of Section 5.3.

3.3.2 Planning Application Ontology

The Plan Application Ontology is used to express the SAV of the mission planner agent. It uses
concepts originally defined by the Planning Domain Definition Language (PDDL). The PDDL
language was originally created by Ghallab et al. (1998) to standardise plan representation.
Concepts are extracted from the language vocabulary and the language grammar is used for
describing the relationships and constraints between these concepts.
For the adaptation mission planning process, the Planning Application Ontology also requires
concepts capable of representing the diagnosis of incidents or problems occurring in some
parts of the mission plan (van der Krogt, 2005). Some of the most important concepts
identified for mission plan adaptability are:
• Resource: state of an object (physical or abstract) in the environment (vehicle, position,
sensor, etc.),
• Action: Modification of the state of resources (calibrate, classify, explore, etc.),
• Gap: A non-executable action,
• Execution: When an action is executed successfully,
• Failure: An unsuccessful execution of an action,
Please note how some of these concepts are also related to concepts of the Core Ontology (e.g.
a list of capability concepts is required to perform a mission action).

4. Adaptive mission planning

The adaptive mission planning process involves the detection of events, the effects that these
events have on the mission plan and the response phase. The detection of events is performed
by the status monitoring agent. The mission plan diagnosis and repair is undertaken by the
adaptive mission planner agent.

4.1 Status Monitor Agent

The Status Monitor Agent considers all symptoms and observations from environmental and
internal data in order to identify and classify events according to their priority and their
nature (critical or incipient). Based on internal events and context information, this agent
is able to infer new knowledge about the current condition of the vehicle with regard to the
availability for operation of its components (i.e. status). In a similar way, environmental data
is also considered for detecting and classifying external events in order to keep the situation
awareness of the vehicle updated.
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 209
11

       
  q−
 1 q

 q−
1 q



Π

0 Π
q−1
 Π̇
q
Π

q Π

q−1
 Π̇
q Π


q

 
ψq−1 
ψ0  ψ

q ψq−1
 ψ

q

  

 
π0 π
q−1
 π

q πq−1
 π

q

Fig. 7. Schematic representation of the autonomous mission generation, replan and repair
processes using partial plan representation of the mission plans.

4.2 Mission plan adaptation agent

This section describes the mission environment model used by the mission plan repair
techniques presented in this chapter. We continue using the mission environment model
previously described. We generalise this model to any step q in the mission execution timeline.
An instance
 ofan UUV mission environment at a given step q can be simply defined as
Πq = Σq , Ωq . The mission domain model Σq contains the set of propositions defining
the available resources in the system PV and the set of actions or capabilities AV . The mission
problem model Ωq contains the current platform state xq and the mission requirements QO .
Based on this model, we analyse how to calculate mission plans on the plan space (Sacerdoti,
1975). A plan space is an implicit directed graph whose vertices are partially specified
plans and whose edges correspond to refinement operations. In a real environment where
optimality can be sacrificed by operability, partial plans are seen as a suitable representation
because they are a flexible constrained-based structure capable of being adapted.
A partial plan ψ is a tuple containing a set of partially instantiated actions and a set of
constraints over these partially grounded actions. Constraints can be of the form of ordering
constraints, interval preservation constraints, point truth constraints and binding constraints.
Ordering constraints indicate the ordering in which the actions should be executed. Interval
preservation constraints link preconditions and effects over actions already ordered. Point
truth constraints assure the existence of precondition facts at certain points of the plan.
Binding constraints on the variables of actions are used to ground the actions to variables
of the domain. Figure 8 shows a partial plan representation of an UUV mission. Partial plans
are flexible to modification. They provide an open approach for handling extensions such as
temporal and resource constraints. Due to nature of the constraints, it is easy to explain a
partial plan to a user. Additionally, it is easily extensible to distributed multi-agent mission
planning.
Figure 7 explains the processes of mission plan repair and mission plan replan for mission
plan adaptation for UUVs using a partial plan representation of the mission plans. At the
initial step, a partial ordered plan ψ0 is generated satisfying the original mission environment
Π0 . The ψ0 is then grounded into the minimal mission plan π0 including all constraints
in ψ0 . At step q, the semantic knowledge-based framework is updated by the diagnosis
information Π̇q providing a modified awareness of the mission environment Πq . From here,
two mission adaptation processes are possible: Mission replan generates a new partial plan ψq ,
as done at the first stage, based only on the knowledge of Πq . On the other hand, mission
plan repair re-validates the original plan by ensuring minimal perturbation of it. Given the
partial plan at the previous step ψq−1 and the diagnosis information Π̇q , the mission repair
problem produces a solution partial plan ψq that satisfies the updated mission problem Πq , by
modifying ψq−1 . The final step for both approaches is to ground ψq to its minimal mission plan
210
12 Autonomous Underwater Vehicles
Underwater Vehicles

Fig. 8. Example of a partial ordered plan representation of an autonomously generated UUV

mission. The ordering constraints are represented using the graph depth, interval
preservation constraints are represented with black arrows, point truth constraints are
represented with PTC-labelled arrows, and binding constraints are shown in the top left box.

πq . It can be seen that mission repair better exploits the orientation capabilities for decision
making: instead of taking the new mission environment as a given, it uses the diagnosis
information about the changes occurred to guide the adaptation process.
We have now identified the benefits of mission plan repair over mission replan. Mission
plan repair modifies the partial plan ψq , so that it uses a different composition, though it
still maintains some of the actions and the constraints between actions from the previous
partial plan. However, mission plan adaptation can also be achieved by mission execution
repair by looking directly at the mission plan instantiation πq . Execution repair modifies
the instantiation of the mission plan πq such that a ground action gqah that was previously
instantiated by some execution eq is newly bound by another action execution instance eq .
Executive repair is less expensive and it is expected to be handled directly by the mission
executive agent. Plan repair, however, is computationally more expensive and requires action
of the mission planner agent.
The objective is to maximise the number of execution repairs over plan repairs and, at the plan
repair level, maximise the number of decisions reused from the previous mission instantiation.
The information provided by the semantic-base knowledge base during the plan diagnosis
phase is critical.
Executive repair fixes plan failures identified in the mission plan during the diagnosis stage.
Our approach uses ontology reasoning in combination with an action execution template to
adapt the mission plan at the executive level.
Once a mission plan πq is calculated by the mission planner, its list of ground actions is
transferred to the executive layer. In this layer, each ground action gqah of πq gets instantiated
into an action execution instance eqt using the action template for the action ah available in
the Core Ontology of the knowledge base. At the end of this phase, each eqt contains the
script of commands required to perform its correspondent ground action. Flexibility in the
execution of an action instance is critical in real environments. This is provided by a timer, an
execution counter, a time-out register and a register of the maximum number of executions in
the action execution instance. Additionally, three different outputs control the success, failure
or time-out of its execution. These elements handle the uncertainty during the execution phase
and enable the executive repair process. This minimise the number of calls to the adaptive
mission planner agent and therefore the response time for adaptation.
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 211
13

Plan repair uses a strategy to repair with new partial plans the plan gaps identified during
the plan diagnosis stage. Our approach uses an iteration of unrefinement and refinement
strategies on a partial-ordered planning framework to adapt the mission plan.
Planning in the plan space is slower than in the state space because the nodes are more
complex. Refinement operations are intended to achieve an open goal from the list of
mission requirements or to remove a possible inconsistency in the current partial plan. These
techniques are based on the least commitment principle, and they avoid adding to the partial
plan any constraint that is not strictly needed. A refinement operation consists of one or more
of the following steps: adding an action, an ordering constraint, a variable binding constraint
or a causal link.
A partial plan is a solution to the planning problem if has no flaw and if the sets
of constraints are consistent. Flaws are either subgoals or threats. Subgoals are open
preconditions of actions that have not been linked to the effects of previous actions. Threats
are actions that could introduce inconsistencies with other actions or constraints. We
implemented a recursive non-deterministic approach based on the Partial ordered Planning
(PoP) framework (Penberthy & Weld, 1992). This framework is sound, complete, and
systematic. Unlike other Plan space planners that handle both types of flaws (goals and
threats) similarly, each PoP recursive step first refines a subgoal and then the associated
threats (Ghallab et al., 2004).
In our implementation, we introduce a previous step capable of performing an unrefinement
of the partial plan when necessary. During the unrefinement strategy we remove refinements
from the partial plan that are reported by the plan diagnosis phase to be affecting the
consistency of the mission plan with the mission environment, i.e. to remove constraints and
finally the actions if necessary.
In simple terms, when changes on the ABox Planning Application Ontology are sensed (Π̇q )
that affect the consistency of the current partial plan ψq−1 , the plan repair process is initiated.
The plan repair stage starts an unrefinement process that relaxes the constraints in the partial
plan ψq−1 that are causing the mission plan to fail.
The remaining temporal mission partial plan ψq −1 is now relaxed to be able to cope with the
new mission environment. However, this relaxation could open some subgoals and introduce
threats in the partial plan that need to be addressed. The plan repair stage then executes
a refinement process searching for a new mission plan ψq that is consistent with the new
mission environment Πq and removing these possible flaws. By doing this, it can be seen that
the new mission plan ψq is not generated again from Πq (re-planned) but recycled from ψq−1
(repaired). This allows re-use of the parts of the plan ψq−1 that were still consistent with Πq .

5. Results
5.1 Architecture
The combination of the status monitor agent, the adaptive mission planner, the mission
executive and the semantic knowledge-based framework is termed as the Semantic-based
Adaptive Mission Planning system (SAMP). The SAMP system implements the four stages
of the OODA-loop. Figure 9 represents the customised version of Figure 3 for SAMP.
The status monitor agent reports to the knowledge base the different changes occurring in
the environment and the modifications of the internal status of the platform. The knowledge
base stores the ontology-based knowledge containing the expert orientation provided a priori
and the observations reported by the status monitor. A mission planner agent generates and
212
14 Autonomous Underwater Vehicles
Underwater Vehicles

Command Action
ALI
Query

Mission Mission
Functional
Executive Planner

Acknowledgment
Notification
Status
Domain Capabilities

Status Knowledge
Monitor Base
Event

Fig. 9. Architecture of the SAMP system. The embedded agents are the planner, executive,
monitor, and knowledge base. These agents interconnect via set of messages. The system
integrates to the functional layer of a generic host platform by an abstract layer interface
(ALI).

adapts mission plans based on the situation awareness stored in the knowledge base. The
mission executive agent executes mission commands in the functional layer based on the
sequence of ground actions received from the mission planner. An Abstract Layer Interface
(ALI) based on JAUS-like messages (SAE, 2008a) over UDP/IP packages implemented using
the OceanSHELL protocol (Oce, 2005) provides independence from the platform’s functional
layer making the system generic and platform independent.

5.2 Simulation results

A set of synthetic simulated scenarios have been implemented to test the performance of the
SAMP system. The tests are based on the mine counter measure (MCM) operation, where
UUVs support and provide solutions for mine-hunting and neutralisation.
A set of 15 selected MCM scenarios were simulated covering the variability of missions
described by the concepts of operations for unmanned underwater vehicles presented in
the UUV (2004) and the JRP (2005). For each scenario, the detection of a failure in one of
the components of the system was simulated. The mission plan was adapted to the new
constraints using replanning methods and the mission plan repair approach based on partial
plans introduced in Section 4.
6
10

1.0
5
10

0.8
4
Time (ms)

10

0.6
Repair
Replan

3
10

0.4
2
10

0.2
1
10

0.0
0
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0.0 0.2 0.4 0.6 0.8 1.0

# Problem Repair
Replan

Fig. 10. Left: A semi-log plot displaying the computational time in miliseconds for replan
(dark grey bars) and repair approaches (light grey bars). Right: Comparison of Plan
Proximity (PP0.5 ) of the replan and repair approaches to the original plan.
The performance of the two approaches was compared by looking at the computation time
and the Plan Proximity (Patrón & Birch, 2009) of the adaptive mission plan provided to the
original reference mission plan. Figure 10 left shows the computation time in milliseconds
required for adapting the mission to the new constraints for replan (dark grey bars) and
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 213
15

repair approaches (light grey bars). Note that a logarithmic scale is used for the time values.
Figure 10 right displays the Plan Proximity to the original plan of the replan strategy result
versus the repair strategy result. It can be seen that plans adapted using the mission repair
strategy tend to be closer to the original plan than using the mission replan strategy. In
these results, 14 out of 15 scenarios were computed faster by using mission plan repair.
This computation was on average 9.1x times faster. Also, 14 out of 15 scenarios showed
that mission plan repair had greater or equal Plan Proximity values as compared to mission
replan. In general, our mission repair approach improves performance and time response
while at the same time finds a solution that is closer to the original mission plan available
before adaptation.

5.3 Experimental results

This section shows the performance of the SAMP system inside the real MCM application
using a REMUS 100 UUV platform (see Fig. 11.left) in a set of integrated in-water field trial
demonstration days at Loch Earn, Scotland (56o 23.1N,4o 12.0W). The REMUS UUV had a
resident guest PC/104 1.4GHz payload computer where the SAMP system was installed.
SAMP was capable of communicating with the vehicle’s control and status modules and
taking control of it by using an interface module that translated the ALI protocol into the
manufacturer’s Remote Control protocol (REM, 2008).

area 1

area 2

Fig. 11. Left: REMUS UUV deployment before starting one of its missions. Right: Procedural
mission uploaded to the vehicle control module and a priori seabed classification information
stored in the knowledge base. The two dark grey areas correspond to the classified seabed
regions.
Figure 11.right shows the procedural waypoint-based mission as it was described to the
vehicle’s control module. This was known as the baseline mission. It was only used to start
the vehicle’s control module with a mission in the area of operation before taking control of
it using the SAMP system. The baseline mission plan consisted on a start waypoint and two
waypoints describing a North to South mission leg at an approximate constant Longitude
(4o 16.2W). This leg was approximately 250 meters long and it was followed by a loiter pattern
at the recovery location. The track obtained after executing this baseline mission using the
vehicle control module is shown in Fig. 11 with a dark line. A small adjustment of the vehicle’s
location can be observed on the top trajectory after the aided navigation module corrects its
solution to the fixes received from the Long Baseline (LBL) transponders previously deployed
in the area of operations.
214
16 Autonomous Underwater Vehicles
Underwater Vehicles

Fig. 12. Core Ontology instances for the demonstration scenario. The diagram represents the
main platform, its components and their capabilities.

Fig. 13. Plan Application Ontology concepts representing the mission planning actions and
their execution parameters and relationships.

On the payload side, the SAMP system was oriented (in the OODA-loop sense) using a
priori information about the environment and the platform and a declarative description
of the goals of the mission. The a priori knowledge and the platform configuration
capabilities was represented using Core Ontology concepts (see Fig. 12). Knowledge about
the environment was provided based on automatic computer-aided seabed classification
information generated from previous existent data (Reed et al., 2006). The two classified
seabed areas are shown in Fig. 11. The declarative description of the mission requirements was
represented using concepts from the Planning Application Ontology. They could be resumed
as ’survey all known areas maximizing efficiency’.

5.3.1 Pre-mission reasoning

Please note that the previously described separation between Core knowledge and Planning
knowledge gracefully aligns with the separation between platform engineers and mission
scientists on current UUV operations. If the platform capabilities were described in Core
Ontology terms by the engineers that manufactured the platform, it can be seen how, by using
the SAMP approach, a scientific operator that only cares about the data should be able to
describe the mission to the platform without knowing anything about the custom properties
of the platform. It is, therefore, important to assist the operator in knowing if the platform
capabilities can match the mission requirements before starting the mission:
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 215
17

• Is this platform configuration suitable to successfully perform this mission?

In order to answer this question, new knowledge could be inferred from the initial Core
Ontology orientation. The Core Ontology rule engine was executed providing with additional
knowledge. A set of predefined rules helped orienting the knowledge base into inferring new
relationships between instances. An example of a rule dealing with the transfer of payload
capabilities to the platform is represented in Eq. 2.

core : isCapabilityO f (?Capability, ?Payload)∧

core : isPayloadO f (?Payload, ?Plat f orm) (2)
→ core : isCapabilityO f (?Capability, ?Plat f orm)
Once all the possible knowledge was extracted, it was possible to query the knowledge base
in order to extract the list of capabilities of the platform (see Eq. 3) and the list of requirements
of the mission (see Eq. 4).

SELECT ?Platform ?Cap WHERE { rdf:type( ?Platform, core:Platform) ∧

(3)
core:hasCapability(?Platform,?Cap) }

SELECT ?Mission ?Req WHERE { plan:hasAction( ?Mission, ?Action) ∧

(4)
plan:hasRequirement( ?Action,?Req ) }

This way, it was possible to autonomously extract that the requirements of the mission of the
experiment were1 :

• core:WaypointManeuver_Capability ∈ jaus:Maneuver_Capability
• core:ComputerAidedClassification_Capability ∈ jaus:Autonomous_RSTA-I_Capability
• core:ComputerAidedDetection_Capability ∈ jaus:Autonomous_RSTA-I_Capability
• core:SidescanSensor_Capability ∈ jaus:Environmental_Sensing_Capability
which were a subset of the platform capabilities. Therefore, for this particular case, the
platform configuration suited the mission requirements.

5.3.2 In mission adaptation

For these experiments, SAMP was given a static location in which to take control of the host
vehicle. At this point, the mission planner agent generated a mission plan based on the
knowledge available and the mission requirements. The instantiation of this mission plan
is described in Fig. 13 using Planning Application Ontology concepts. The mission was then
passed to the executive agent that took control of the vehicle for its execution.
While the mission was executed the status monitor agent maintained the knowledge
base updated (in the OODA-loop sense) by reporting changes in the status of hardware
components, such as batteries and sensors, and external parameters, such as water currents.
When observations indicated that some of these changes were affecting the mission under
execution, the mission planner was activated in order to adapt the mission to the changes.
This indication was detected by the planner agent by querying the knowledge base with the
following question:
1 RSTA-I i.e., Reconnaissance/Surveillance/Target Acquisition & Identification capability concepts
inherited from JAUS (SAE, 2006)
216
18 Autonomous Underwater Vehicles
Underwater Vehicles

0

50

North (m)(m)

100
North
Depth
Depth

150

200

250

100
50 0 50 100 150 200

North East
East (m)
East (m) North
East

Fig. 14. Left: Vehicle’s track during mission in a North-East coordinate frame projection with
the origin at the starting point of the mission. Right: Three-dimensional display of the
vehicle’s track during the mission (Note that depth coordinates are not to scale).

• Are the observations coming from the environment affecting the mission currently under execution?
In order to explain the reasoning process involved during the event detection, diagnosis and
response phases of the mission adaptation process, a component fault as an internal event
was temporarily simulated in the host vehicle. The fault simulated the gains of the starboard
transducer of the sidescan sonar dropping to their minimum levels half way through the lawn
mower survey of the first area.
For the detection phase, the low gain signals from the transducer triggered a symptom
instance, which had an associated event level. This event level, represented in the Status
Monitoring Application Ontology using a value partition pattern, plays a key role in the
classification of the instances in the Event concept between being critical or incipient. This
classification is represented axiomatically in the Eqs. 5 and 6.

status:CriticalEvent  status:Event status:causedBySymptom . . .

(status:Symptom status:hasEventLevel . . . (5)
(status:Level status:High))

status:IncipientEvent  . . .
status:Event status:causedBySymptom . . .
(6)
(status:Symptom status:hasEventLevel . . .
(status:Level status:Med))

After the Event individuals were re-classified, the Status property of the related component in
the Core Ontology was updated.
During the diagnosis event phase, a critical status of a component is only considered to be
caused by a critical event. Therefore, due to the fact that the sidescan sonar component is
composed of two transducers, port and starboard, one malfunctioned transducer was only
diagnosed as an incipient Status of the overall sidescan sonar component.
During the response phase, the Status property of the Core Ontology components were used
by the mission planner to perform the plan diagnosis of the mission under execution. The
query to the knowledge base shown in Eq. 7 reported that the two survey actions in the
mission plan were affected by the incipient status of the sidescan sonar.
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 217
19

a)

b)

c)

d)

e)

Fig. 15. Vehicle telemetry (top to bottom): a) vehicle velocity (m/s), b) compass heading
(degrees), c) altitude (m), d) depth (m) and e) reconstructed profile of the seabed bathymetry
(m) during the mission, all plotted against mission time (s).

SELECT ?Mission ?Action ?Param ?Status

WHERE { plan:hasAction( ?Mission, ?Action) ∧
(7)
plan:hasExecParam( ?Action,?Param) ∧
plan:hasStatus( ?Param, ?Status) }
An incipient Status of the action parameters indicates that the action can still be performed by
adapting the way it is being executed, an execution repair. If both transducers were down, a
critical status of the sidescan sensor is diagnosed and a plan repair adaptation of the mission
plan would have been required instead. In that case, the adaptive mission planner would have
looked for redundant components or similar capabilities to perform the action or to drop the
action from the plan.
The same procedure was used after the transducer recovery was reported to adapt the survey
action to the normal pattern during the second lawn mower survey. In a similar process,
SAMP adapted the lawnmower pattern survey of the areas to the detected water current Status
at the moment of initialising the survey of the areas.
The timeline of the mission executed using the SAMP approach is described in the following
figures: Figure 14 represents the final trajectory of the vehicle in 2D and 3D using a
North-East coordinate frame projection with the origin at the starting point of the mission.
Figure 15 displays the vehicle’s telemetry recorded during the mission. It includes vehicle’s
velocity, compass heading, altitude, depth measurements and processed bathymetry over
time. Figure 16 shows subset of the variables being monitored by the status monitor agent
that were relevant to this experiment. These variables include direction of water current,
remaining battery power, the availability of the transducers in the sidescan sensor and the
mission execution status. Figure 17 represents the system activity of the payload computer
recorded during the mission. The system activity logs show percentage of processor usage,
memory usage, network activity and disk usage.
218
20 Autonomous Underwater Vehicles
Underwater Vehicles

100 200 300

a)

0
0 500 1000 1500 2000

0.8
b)

0.4
●

0.0
0 500 1000 1500 2000

c)

d)

e)

f)

Fig. 16. Status monitoring (top to bottom): a) direction of water current (degrees), b) speed of
water current (m/s), c) battery power (Wh), d) sidescan sensor port and e) starboard
transducers availability (on/off) and f) mission status binary flag, all plotted against mission
time (s).
20

●
15
10

a)
5
0

0 500 1000 1500 2000

b)

c)

d)

Fig. 17. System activity (top to bottom): a) % processor usage, b) % memory usage, c)
network activity (packets/s) and d) disk usage (I/O sectors/s), all plotted against mission
time (s).

Each of the symbols , , ♦, and  on the aforementioned figures represents a point

during the mission where an event occurred. Symbol represents the point where SAMP
takes control of the vehicle. Note a change on the host platform mission status binary flag that
becomes 0x05, i.e. the mission is active (0x01) and the payload is in control (0x04) (Figure 16.e).
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 219
21

Also, a peak on the CPU usage can be noted as this is the point where the mission partial plan
gets generated (Figure 17.a).
Symbol  represents the point where the vehicle arrives to perform the survey of the area. At
this point, the action survey gets instantiated based on the properties of the internal elements
and external factors. Although the Loch waters where the trials were performed were very still
(see Figure 16.b), note how the vehicle heading during the lawnmower pattern performed to
survey the areas follows the water current direction sensed at the arrival (approx. 12o , Symbol
 - Figure 16.a) in order to minimize drag and maximise battery efficiency. The heading of the
vehicle during the survey can be observed in Figure 14 and Figure 15.b. The link between
the vehicle heading in relation to the water current direction and its effect on the battery
consumption was expert orientation knowledge captured by a relationship property between
the two concepts in the Core Ontology.
Symbol ♦ represents the point when the status monitor agent detects and reports a critical
status in the starboard transducer of the sidescan sonar (Figure 16.d). It can be seen how the
lawnmower pattern was adapted to cope with the change and to use the port transducer to
cover the odd and even spacing of the survey. This pattern avoids gaps in the sidescan data
under the degraded component configuration and maximises sensor coverage for the survey
while the transducer is down.
Symbol indicates the point where the starboard transducer recovery is diagnosed. It can
be observed how the commands executing the action are modified in order to optimise the
survey pattern and minimise distance travelled. Although also being monitored, the power
status does not report any critical status during the mission that requires modification of the
actions (Figure 16.c).
Symbol  shows the location where all the mission goals are considered achieved and the
control is given back to the mission control of the host vehicle (see Symbol  - Figure 16.e
shows the mission is still active but the payload is not longer in control (0x01) ). From this
point the host vehicle’s control module takes the control back and drives the vehicle to the
loiter at the recovery location.

6. Conclusion and future work

The underwater domain is a challenging environment in which to maintain the operability of
an UUV. Operability can be improved with the embedded adaptation of the mission plan.
We implement a system capable of adapting mission plans autonomously in the face of
events while during a mission. We do this by using a combination of ontological hierarchical
representation of knowledge and adaptive mission plan repair techniques. The advantage
of this approach is that it maximises robustness, system performance and response time.
The system performance has been demonstrated in simulation. Additionally, the mission
adaptation capability is shown during an in-water field trial demonstration.
In our fully integrated experiments we achieved the following:
• Knowledge based framework: We have presented a semantic-based framework that provides
the core architecture for knowledge representation for service oriented agents in UUVs.
The framework combines the initial expert orientation and the observations acquired
during mission in order to improve the situation awareness in the vehicle. This is currently
unavailable in UUVs.
• Goal-oriented plan vs. waypoint-based plan: The system uses a goal-oriented approach in
which the mission is described in terms of ’what to do’ instead of a ’how to do’ it. The
220
22 Autonomous Underwater Vehicles
Underwater Vehicles

mission is parametrised and executed based on the available knowledge and vehicle
capabilities. This is the first time that an approach to goal-based planning is applied to
the adaptation of an underwater mission in order to maintain platform’s operability.
• Adaptation to environmental parameters and internal issues: The approach shows adaptability
to environmental elements, such as water current flows in order to improve mission
performance. The approach is also capable of dealing with the critical status of certain
components in the platform and can react accordingly.
• Platform agnostic: The approach is platform independent making it readily applicable to
other domains, such as ground or air vehicles.
SAMP is open to event detections coming from other embedded service-oriented agents. We
are planning to apply the approach to more complex scenarios involving other embedded
agents, such as agents for automatic target recognition. We are also planning to extend it to
a team of vehicles performing a collaborative mission. In this scenario, agents are distributed
across the different platforms. We are currently working towards a shared situation awareness
for a team of vehicles to which every team member possess the awarenessrequired for its
responsibilities. The main advantage of our semantic-based approach is the low bandwidth
required to share id-coded ontological concepts and, therefore, to cope with the underwater
acoustic communication limitations.

7. References
Adams, J. A. (2007). Unmanned vehicle situation awareness: A path forward, Proceedings of
the 2007 Human Systems Integration Symposium, Annapolis, Maryland, USA.
Arkin, R. C. (1998). Behavior-based Robotics, Massachusetts Institute of Technology Press, ISBN:
978-0262011655.
Bellingham, J., Consi, T., Beaton, R. & Hall, W. (1990). Keeping layered control simple,
Proceedings of the (1990) Symposium on Autonomous Underwater Vehicle Technology
(AUV’90), pp. 3–8.
Bellingham, J., Kirkwood, B. & Rajan, K. (2006). Tutorial on issues in underwater robotic
applications, 16th International Conference on Automated Planning and Scheduling
(ICAPS’06).
Benjamin, M. R., Leonard, J. J., Schmidt, H. & Newman, P. M. (2009). An overview of
MOOS-IvP and a brief users guide to the IvP Helm autonomy software, Technical
Report MIT-CSAIL-TR-2009-028, Computer Science and Artificial Intelligence Lab,
MIT.
Bennet, A. A. & Leonard, J. J. (2000). A behavior-based approach to adaptive feature
detection and following with autonomous underwater vehicles, IEEE Journal of
Oceanic Engineering 25(2): 213–226.
Blomqvist, E. & Sandkuhl, K. (2005). Patterns in ontology engineering âĂŞ classification
of ontology patterns, 7th International Conference on Enterprise Information Systems,
Miami, USA.
Boyd, J. (1992). A discourse on winning and losing - organic design for command and control,
Technical report, http://www.d-n-i.net/dni/john-r-boyd/.
Brooks, R. (1986). A robust layered control system for a mobile robot, Journal of Robotics and
Automation RA-2(1): 14–23.
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 221
23

Caccia, M. & Veruggio, G. (2000). Guidance and control of a reconfigurable unmanned

underwater vehicle, Control Engineering Practice 8(1): 21 – 37.
Carreras, M., Palomeras, N., Ridao, P. & Ribas, D. (2007). Design of a mission control system
for an AUV, International Journal of Control 80(7): 993–1007.
Carreras, M., Ridao, P., Garcia, R. & Batlle, J. (2006). Behaviour Control of UUVs, G. Roberts and
R. Sutton, Eds Advances in Unmanned Maritime Vehicles (IEE Control Engineering),
ISBN: 978-0863414503, chapter 4, pp. 67–86.
Chen, H., Perich, F., Finin, T. & Joshi, A. (2004). Soupa: Standard ontology for ubiquitous
and pervasive applications, International Conference on Mobile and Ubiquitous Systems:
Networking and Services, Boston, MA.
Farrell, J. A., Pang, S. & Li, W. (2005). Chemical plume tracing via an autonomous underwater
vehicle, IEEE Journal of Oceanic Engineering 30, 2: 428–442.
Fossen, T. I. (1994). Guidance and Control of Ocean Vehicles, John Wiley and Sons, ISBN:
0-471-94113-1.
Ghallab, M., Howe, A., Knoblock, C., McDermott, D., Ram, A., Veloso, M., Weld, D. & Wilkins,
D. (1998). PDDL: The planning domain definition language, Technical report, Yale
Center for Computational Vision and Control.
Ghallab, M., Nau, D. & Traverso, P. (2004). Automated Planning: Theory and Practice, Morgan
Kaufmann, ISBN: 1-55860-856-7.
Griffiths, G. (2005). AUTOSUB under ice, Ingenia, ISSN: 1472-9768 22: 30–32.
Gruber, T. R. (1995). Towards principles for the design of ontologies used for knowledge
sharing, International Journal Human-Computer Studies 43: 907–928.
Hagen, P. E. (2001). AUV/UUV mission planning and real time control with the HUGIN
operator systems, Proceedings of the IEEE International Conference Oceans (Oceans’01),
Vol. 1, Honolulu, HI, USA, pp. 468–473.
Jakuba, M. (2007). Stochastic Mapping for Chemical Plume Source Localization with Application
to Autonomous Hydrothermal Vent Discovery, Doctor of philosophy in mechanical
engineering, MIT/WHOI Joint Program in Applied Ocean Physics and Engineering.
JRP (2005). Joint robotics program master plan FY2005, Technical report, US Department of
Defense.
Kokar, M. M., Matheus, C. J. & Baclawski, K. (2009). Ontology-based situation awareness,
Information Fusion 10(1): 83–98.
Matheus, C., Kokar, M. & Baclawski, K. (2003). A core ontology for situation awareness,
Proceedings of the Sixth International Conference of Information Fusion 1: 545 – 552.
McGann, C., Py, F., Rajan, K. & Henthorn, R. (2008). Adaptive control for autonomous
underwater vehicles, Association for the Advancement of Artificial Intelligence (AAAI’08),
Chicago, IL, USA.
McGann, C., Py, F., Rajan, K., Henthorn, R. & McEwen, R. (2008). A deliberative architecture
for AUV control, International Conference on Robotic and Autonomation (ICRA’08),
Pasadena, CA, USA.
McGann, C., Py, F., Rajan, K. & Olaya, A. G. (2009). Integrated planning and execution for
robotic exploration, International Workshop on Hybrid Control of Autonomous Systems,
Pasadena, CA, USA.
McGann, C., Py, F., Rajan, K., Thomas, H., Henthorn, R. & McEwen, R. (2007). T-REX: A
model-based architecture for AUV control, Workshop in Planning and Plan Execution
222
24 Autonomous Underwater Vehicles
Underwater Vehicles

for Real-World Systems: Principles and Practices for Planning in Execution, International
Conference of Autonomous Planning and Scheduling (ICAPS’07).
Miguelanez, E., Lewis, R., K.Brown, Roberts, C. & Lane, D. (2008). Fault diagnosis of a
train door system based on the semantic knowledge representation, The 4th IET
International Conference on Railway Condition Monitoring RCM 2008, Derby Conference
Centre, Derby, UK, pp. 1 –6.
Miguelanez, E., Patrón, P., Brown, K., Petillot, Y. R. & Lane, D. M. (2010). Semantic
knowledge-based framework to improve the situation awareness of autonomous
underwater vehicles, IEEE Transactions on Knowledge and Data Engineering (In Press)
PP(99).
Newman, P. (2002). MOOS – a mission oriented operating suite, Technical report, Department
of Ocean Engineering, Massachusetts Institute of Technology.
Niles, I. & Pease, A. (2001). Towards a standard upper ontology, Proceedings of the 2nd
International Conference on Formal Ontology in Information Systems (FOIS-2001) .
Oce (2005). OceanSHELL: An embedded library for distributed applications and
communications, Technical report, Ocean Systems Laboratory, Heriot-Watt University.
Oliveira, P., Pascoal, A., Silva, V. & Silvestre, C. (1998). Mission control of the MARIUS AUV:
System design, implementation, and sea trials, International Journal of Systems Science
29(10): 1065–1085.
Palomeras, N., Ridao, P., Carreras, M. & Silvestre, C. (2009). Using petri nets to specify and
execute missions for autonomous underwater vehicles, Proceedings of the IEEE/RSJ
International Conference on Intelligent Robots and Systems (IROS’09).
Patrón, P. (2009). Embedded knowledge and autonomous planning, Sea Technology Magazine,
ISSN. 0093-3651 50(4): 101.
Patrón, P. & Birch, A. (2009). Plan proximity: an enhanced metric for plan stability, Workshop
on Verification and Validation of Planning and Scheduling Systems, 19th International
Conference on Automated Planning and Scheduling (ICAPS’09), Thessaloniki, Greece,
pp. 74–75.
Patrón, P., Evans, J. & Lane, D. M. (2007). Mission plan recovery for increasing vehicle
autonomy, Proceedings of the Conference of Systems Engineering for Autonomous Systems
from the Defence Technology Centre (SEAS-DTC’07), Edinburgh, UK.
Patrón, P., Lane, D. M. & Petillot, Y. R. (2009a). Continuous mission plan adaptation for
autonomous vehicles: balancing effort and reward, 4th Workshop on Planning and Plan
Execution for Real-World Systems, 19th International Conference on Automated Planning
and Scheduling (ICAPS’09), Thessaloniki, Greece, pp. 50–57.
Patrón, P., Lane, D. M. & Petillot, Y. R. (2009b). Interoperability of agent capabilities for
autonomous knowledge acquisition and decision making in unmanned platforms,
Proceedings of the IEEE International Conference Oceans Europe (Oceans Europe’09),
Bremen, Germany.
Patrón, P., Miguelanez, E., Cartwright, J. & Petillot, Y. R. (2008). Semantic knowledge-based
representation for improving situation awareness in service oriented agents of
autonomous underwater vehicles, Proceedings of the IEEE International Conference
Oceans (Oceans’08), Quebec, Canada.
Patrón, P., Miguelanez, E., Petillot, Y. R. & Lane, D. M. (2008). Fault tolerant adaptive mission
planning with semantic knowledge representation for autonomous underwater
Embedded Knowledge and Autonomous Planning:
The Path
Embedded Towards
Knowledge Permanent
and Autonomous Presence
Planning: the of Underwater
Path Towards Permanent Networks
Presence of Underwater Networks 223
25

vehicles, Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and

Systems (IROS’08), Nice, France, pp. 2593–2598.
Patrón, P. & Petillot, Y. R. (2008). The underwater environment: A challenge for planning,
Proceedings of the Conference of the UK Planning Special Interest Group (PlanSIG’08),
Edinburgh, UK.
Penberthy, J. & Weld, D. S. (1992). UCPOP: A sound, complete, partial order planner for ADL,
Proceedings of the National Conference on Knowledge Representation and Reasoning (KR),
pp. 103–114.
Portele, C. (2007). OpenGISÂő geography markup language (GML) encoding standard v.3.2.1,
Technical report, Open Geospatial Consortium Inc.
Rajan, K., McGann, C., Py, F. & Thomas, H. (2007). Robust mission planning using
deliberative autonomy for autonomous underwater vehicles, Workshop in Robotics
in challenging and hazardous environments, International Conference on Robotics and
Automation (ICRA’07).
Rajan, K., Py, F., McGann, C., Ryan, J., O’Reilly, T., Maughan, T. & Roman, B. (2009). Onboard
adaptive control of AUVs using automated planning and execution, International
Symposium on Unmanned Untethered Submersible Technology (UUST’09), Durham, NH,
USA.
Reed, S., Ruiz, I., Capus, C. & Petillot, Y. (2006). The fusion of large scale classified side-scan
sonar image mosaics, IEEE Transactions on Image Processing 15(7): 2049–2060.
REM (2008). REMUS remote control protocol v1.12, Technical report, Hydroid Inc.
Ridao, P., Batlle, J., Amat, J. & Roberts, G. (1999). Recent trends in control architectures
for autonomous underwater vehicles, International Journal of Systems Science
30(9): 1033–1056.
Rosenblatt, J. K., Williams, S. B. & Durrant-Whyte, H. (2002). Behavior-based control for
autonomous underwater exploration, International Journal of Information Sciences,
145(1–2): 69–87.
Sacerdoti, E. (1975). The nonlinear nature of plans, International Joint Conference on Artificial
Intelligence (IJCAI’75), pp. 206–214.
SAE (2006). Society of automotive engineers AS-4 AIR5664 JAUS history and domain model,
Technical report, SAE International Group.
SAE (2008a). Society of automotive engineers AS-4 AIR5665 JAUS architecture framework for
unmanned systems, Technical report, SAE International Group.
SAE (2008b). Society of automotive engineers AS-4 ARP6012 JAUS compliance and
interoperability policy, Technical report, SAE International Group.
Sirin, E., Parsia, B., Grau, B. C., Kalyanpur, A. & Katz, Y. (2007). Pellet: A practical owl-dl
reasoner, Web Semantics: Science, Services and Agents on the World Wide Web 5(2): 51–53.
Turner, R. (2005). Intelligent mission planning and control of autonomous underwater
vehicles, Workshop on Planning under uncertainty for autonomous systems, 15th
International Conference on Automated Planning and Scheduling (ICAPS’05).
Turner, R. M. (1995). Context-sensitive, adaptive reasoning for intelligent auv control: Orca
project update, In Proceedings of the Ninth International Symposium on Unmanned,
Untethered Submersible Technology (UUST’95), Durham, NH, USA, pp. 426–435.
UUV (2004). The navy unmanned undersea vehicle (UUV) master plan, Technical report, US
Department of the Navy.
224
26 Autonomous Underwater Vehicles
Underwater Vehicles

van der Krogt, R. (2005). Plan repair in single-agent and multi-agent systems, PhD thesis,
Netherlands TRAIL Research School.
van Heijst, G., Schreiber, A. & Wielinga, B. (1996). Using explicit ontologies in kbs
development, International Journal of Human-Computer Studies 46(2-3).
von Alt, C. (2010). Remus technology 2020 and beyond, NATO Naval Armaments Group –
Maritime Capability Group 3 on Mines, Mine Countermeasures and Harbour Protection –
Industry Day, Porton Down, UK.
Yuh, J. (2000). Design and control of autonomous underwater robots: A survey, Journal of
Autonomous Robots 8(1): 7–24.
 10

Deep-Sea Fish Behavioral Responses to

Underwater Vehicles: Differences Among
Vehicles, Habitats and Species
Franz Uiblein
Institute of Marine Research, Bergen, Norway and South African Institute of Aquatic
Biodiversity, Grahamstown,
South Africa

1. Introduction
Fishes have a wide range of perceptual capabilities allowing them to behaviorally respond
to various environmental stimuli such as visual, acoustic, mechanical, chemical, and
electromagnetic signals. In our “noisy” world of today many artificially evoked signals pass
through aquatic habitats, where fishes perceive them and respond to in often unpredictable
manner. Proper distinction between natural and artificially evoked (=”disturbed”) behavior
is of utmost importance in ecological studies that try to identify the prevailing factors and
mechanisms influencing fish abundance, distribution and diversity.
As we know today, the need to consider human-induced behavioral disturbance as an
important factor in ecological studies (Beale 2007) applies even to inhabitants of remote
aquatic habitats such as the deep sea. In situ studies using various types of underwater
vehicles (UV’s) have significantly changed the conception that the inhabitants of the deep,
dark and mostly cold ocean are less behaviorally active and hence less susceptible to
anthropogenic disturbance. While direct observation of deep-sea animals goes back to the
time of William Beebe in the 1930s, in situ studies of deep ocean organisms and their
habitats have become increasingly more common during the last 50 years.
After initial use for exploration and discovery of yet unknown habitats and organisms, UV’s
were adopted to systematically investigate the ecology of deep-sea organisms, especially the
larger and easier observable fauna in the open water and close to the bottoms. In analogy to
census studies conducted by divers in shallow waters, vertical or horizontal transects with
underwater vehicles were used to obtain density or distributional data of fishes (e.g.,
Yoklavich et al. 2007, Uiblein et al. 2010). Distinct fish species or closely related taxonomic
groups were found to occur at relatively high densities during such transects allowing
quantitative behavioral investigations.
Early in situ exploration encountered first evidence of pelagic and bottom-associated
(demersal) fishes living at depths well below 200 m being behaviorally active similar to
shallow-water species (Beebe 1930, Heezen & Hollister 1971). These preliminary behavioral
observations were followed by detailed studies of locomotion behavior and habitat
utilization based mainly on video equipment employed during bottom transects with
manned submersibles (e.g., Lorance et al. 2002, Uiblein et al. 2002, 2003) and later with
226 Autonomous Underwater Vehicles

ROV’s (e.g., Trenkel et al. 2004a, Lorance et al. 2006). Quantitative behavioral comparisons
conducted with the submersible Nautile clearly showed that fish species differ among each
other in the way they swim and in their vertical positioning above the bottom (Uiblein et al.
2003). Moreover, distinct responses to the approaching vehicle were identified which
needed to be analyzed in detail so to be able to distinguish natural behavior from responses
to anthropogenic disturbance. That underwater vehicles have a disturbance effect on fish
behavior has also important consequences for fish density calculations from in situ transects,
as the data may not reflect natural conditions when disturbance responses are intense
and/or occur frequently (Trenkel et al. 2004b, Stone et al. 2008).
Disturbance responses in deep-sea fishes may be caused by a number of factors like noise
produced by motors and thrusters, light used for illumination purposes, motion,
electromagnetic fields, or odor plumes deriving from the vehicle. Detailed investigations
regarding the actual source(s) of disturbance are generally lacking. Here, a description and
categorization of disturbance responses is provided and differences between vehicles,
habitats, and species are elaborated. These data suggest that disturbance responses are
manifold and can – by themselves – reveal interesting insights into the life modes of deep-
sea fishes. In addition, when disturbance responses are identified, natural behavior (e.g.,
locomotion and vertical positioning above bottom) can be filtered out and studied
independently of artificial evocation.
Here, nine case studies based on manned submersible and ROV video transects in the
deep North Atlantic are presented dealing subsequently with differences in disturbance
responses between underwater vehicles, dive transects (habitats), and co-occurring
species/species groups. In addition, a separate section is devoted to combined analyses of
natural behavior and disturbance responses, to show the full picture. These results are
discussed referring to (1) novel insights about deep-sea fish artificially and naturally
aroused behavior, (2) the need for consideration and integration of all influential factors in
the behavioral analysis and interpretation, and (3) future technological possibilities and
challenges towards optimizing in-situ investigations on the behavior and ecology of deep-
sea fishes.

2. Material and methods

Video recordings from the areas of the Bay of Biscay and the northern Mid-Atlantic Ridge
made during six dives with four different underwater vehicles were studied. The
underwater vehicles were as follows (Table 1): the manned submersible Nautile and the
ROV Victor 6000 (both at IFREMER, www.ifremer.fr), the ROV Aglantha (IMR,
www.imr.no), and the ROV Bathysaurus (ARGUS, www.argus-rs.no). Each dive consisted
of one to three horizontal transects close to the bottom which lasted between 10 and 174
minutes and covered various depth ranges between 812 and 1465 m (Table 1). During
transects the respective vehicle moved slowly (ca. 0.5 to 1.0 knots on average) above the
bottom, mostly in straight lines, sometimes interrupted by short stops.
Each of the 10 total transects crossed a distinct habitat within canyons or deep-sea terraces of
the Bay of Biscay (Nautile, Victor 6000) and slopes or valleys of the northern Mid-Atlantic
Ridge (Aglantha, Bathysaurus) (Table 1). The Mid-Atlantic Ridge study area was divided in
a southern investigation box, close to the Azores, and a northern box situated in the area
south and north of the Charlie Gibbs Fracture Zone (Table 1).
Deep-Sea Fish Behavioral Responses to Underwater Vehicles:
Differences Among Vehicles, Habitats and Species 227

Table 1. Overview of dives, vehicles and video transects, with numbers of encountered fish
per transect. Samples analyzed are highlighted. For further explanations see text.

Behaviour Category
No Close Far Arriving
Disturbance response
response distance distance disturbed
Vertical position in water Well above
Close to bottom Far above bottom
column bottom
Locomotion behaviour Inactive Drifting Station holding Forward moving
Table 2. Overview of the behavioral categories studied
The four species/species groups selected for detailed analysis were the roundnose grenadier
Coryphaenoides rupestris (family Macrouridae; Fig.1), the orange roughy Hoplostethus
atlanticus (family Trachichthyidae; Fig.1) the false boarfish Neocyttus helgae (family
Oreosomatidae; Fig.1) and codling (family Moridae). The term “codling” includes the most
common Lepdion eques (North Atlantic codling; Fig. 1), its congeners L. guentheri and L.
schmidti, and the slender codling Halagyreus johnssonii. Identification of species/species
groups was based on the size and form of the body, head and fins, and color patterns and
distributional data from the respective area deriving from collected material.
The recording of all behaviors started immediately after a fish appeared on the video screen.
Four main behaviors, overall activity level, disturbance response, locomotion, and vertical
positioning above the bottom, each consisting of two or more categories, were recorded for
subsequent statistical analysis (Table 2). Fishes visualized on video with high or increasing
swimming speed indicating burst swimming in response to prior disturbance by the
submersible (“arriving disturbed”) were excluded from further-going behavioral analyses.
During the subsequent behavioral recordings, the UV frequently got closer to the fishes,
with increasing illumination intensity caused by the front lights. If a disturbance response
was observed during this process (i.e. a marked change in activity level and/or locomotion
behavior), the recordings of locomotion or vertical body positioning were stopped
immediately before the occurrence of this behavioral change. The disturbance response
during UV approach was split into two separate categories, depending if it happened still at
far distance or at close distance to the UV and mostly within the highest illumination radius.
228 Autonomous Underwater Vehicles

Fig. 1. Photographs of studied fish species (North Atlantic codling was the most common
species of the codling group)
For the analysis of undisturbed natural behavior, four locomotion activity categories were
identified: “inactive” (Table 2) (= without any movement), “station holding” (= body
stationary with active swimming against current), “drifting” (= movement in lateral or
backward direction with or without swimming activity), and “forward movement” (= clear
active forward swimming movements). Three categories for vertical body positioning in
relation to the bottom surface were determined: “close to bottom” (= positioned at the
bottom or at distances of less than one body length above the bottom), “well above bottom”
(= distance from bottom exceeds one body length), and “far above bottom” (= distance from
bottom exceeds three body lengths).
In order to reduce the number of influential factors comparisons between underwater
vehicles and species/species groups were mostly restricted to the same transect or area and
comparisons among habitats were restricted to single species. Only samples with 19 or more
individuals per species/species group encountered per transect were analyzed to allow
statistical comparisons in all instances. For statistical comparisons of categorical data among
species/species groups and habitats, G-tests of independency were carried out (Sokal &
Rohlf 1981).
Deep-Sea Fish Behavioral Responses to Underwater Vehicles:
Differences Among Vehicles, Habitats and Species 229

3. Results
The behavioral data of 501 fishes from the four selected species/species groups were
analyzed. Apart from a single exception (codling in dive transect OB22-1) disturbance
responses occurred during all transects and in all species/species groups. On average 44 %
of all fishes showed disturbance and in 7 of the 15 total observational sets (= species-transect
combinations) that were analyzed, more than 50 % of the fish displayed disturbance
responses. While pre-arrival disturbance was relatively rare (14 % of all disturbed behavior
registered), disturbance responses at far distance occurred most frequently (59 %). The
disturbance responses were only rarely directed towards any of the four UV’s used. No clear
signs of attraction or aggressive responses triggered by the UV’s could be observed in any of
the four species/species groups.
Differences between underwater vehicles (Fig.2)
The codling showed a significant difference (p<0.005) in disturbance responses between two
dive transects performed in the same area at the Mériadzek terrace, Bay of Biscay, one with
the manned submersible Nautile (transect OB22-1, Table 1) and the other with the ROV
Victor 6000 (transect VT-1, Table 1). While no disturbance response was registered during
the dive with Nautile, 35 % of the individuals encountered during the ROV transect showed
clear signs of disturbance. Among the disturbed fish 23 % showed pre-arrival disturbance,
while 54 % responded at far distance and 23 % responded at short distance to the
approaching vehicle. Regarding undisturbed natural behavior, no significant differences in
both vertical positioning and locomotion behavior were found between the two transects.

Codling
100
No response
Close distance
80 Far distance
Arriving disturbed
% of Activity

60

40

20

0
OB22-1 VT1-2

Fig. 2. Disturbance responses of codling during a manned submersible transect (left) and a
ROV transect (right) in the area of Mériadzek Terrace, Bay of Biscay
Differences between dive transects and habitats (Fig. 3)
Orange roughy showed significant differences in disturbance responses (p<0.01; Fig. 3a)
between two transects that crossed adjacent habitats at similar depths (812-879 m) during
dive ME10 (Table 1) with the ROV Aglantha on the northern Mid-Atlantic Ridge, just south
230 Autonomous Underwater Vehicles

of the Charlie Gibbs Fracture zone. Each of the three categories of disturbance responses
decreased in frequency between the first and the second transect thus indicating less
responsiveness. Both vertical positioning and locomotion behavior did not differ
significantly between transects.

Orange roughy Codling

a 100 b 100
No response No response
Close distance Close distance
80 Far distance 80 Far distance
Arriving disturbed Arriving disturbed
% of Activity

% of Activity
60 60

40 40

20 20

0 0
ME10-2 ME10-3 VT1-1 VT1-2

Fig. 3. Disturbance responses of (a) orange roughy during two subsequent ROV transects in
the area of the northern Mid-Atlantic Ridge and (b) codling during two subsequent ROV
transects in the area of Mériadzek Terrace, Bay of Biscay
The codling showed a significant decrease in disturbance responses (p<0,005; Fig. 3b)
between the first and second transect of ROV dive VT1 (Table 1) on the Mériadzek Terrace,
Bay of Biscay. These two transects covered different depth zones (1392-1454 vs. 1208-1228
m), the first (VT1-1) being clearly deeper. Neither vertical positioning nor locomotion
behavior differed significantly between the two transects.
Differences between co-occurring species/species groups (Fig. 4)
During the manned submersible transect OB22-1 on the Mériadzek terrace, roundnose
grenadier differed significantly in disturbance responses (p<0.0001; Fig. 4a) from the
codling. The former showed all three categories of disturbance, while the latter showed no
disturbance responses at all (see also first case study; Fig. 2). Regarding natural behaviour,
no differences in vertical positioning occurred, but roundnose grenadier showed
significantly more forward movement and less station holding than codling (p<0.01).
During ROV dive transect VT1-1 the codling and the boarfish differed significantly from
each other in disturbance responses (p<0.005; Fig. 4b) with the boarfish showing clearly less
disturbed arrival and close-distance responses to the approaching vehicle. At far distance
from the ROV, the frequency of disturbance responses was similar in both taxa. In addition,
significant differences occurred both in vertical positioning and locomotion behavior which
are dealt with at the end of the next section.
Variation in natural behavior and disturbance responses
Four different comparative data sets were selected (1) to exemplify situations with
disturbance responses occurring at constant or variable rates between transects/habitats or
between species/species groups and (2) to analyze in detail the undisturbed, natural vertical
Deep-Sea Fish Behavioral Responses to Underwater Vehicles:
Differences Among Vehicles, Habitats and Species 231

OB22-1 VT1-1
a 100 b 100
No response No response
Close distance Close distance
80 Far distance 80 Far distance
Arriving disturbed Arriving disturbed
% of Activity

% of Activity
60 60

40 40

20 20

0 0
Codling Roundnose Codling False
grenadier boarfish
Fig. 4. Disturbance responses of (a) codling and roundnose grenadier during a manned
submersible transect and (b) codling and false boarfish during a ROV transect in the area of
Mériadzek Terrace, Bay of Biscay
positioning and locomotion behavior of these species/species groups during the same dive
transects.
a. Differences only in locomotion behavior (Fig. 5)
The disturbance responses of codling did not differ significantly between two transects
(ME10-1, ME10-3, Table 1) of a dive with the ROV Bathysaurus on the Mid Atlantic Ridge
(Fig. 5a). There was however a significant difference in locomotion behavior (p<0.05).
During the first transect all individuals encountered were active and mostly station holding,
while several were inactively sitting on the bottom during the third transect, with less fish
station holding. Drifting and forward moving occurred in both transects at rather similar
rates. No significant differences in vertical positioning occurred.
During the ROV transect VT3-2 in the Bay of Biscay roundnose grenadier and codling did
not differ significantly from each other in disturbance response and vertical positioning (Fig.
5b). However, they clearly differed in locomotion behavior (p<0.005), with roundnose
grenadier showing less frequently station holding and more often drifting and forward
movement than the codling.
b. Differences only in locomotion and vertical positioning (Fig. 6a)
In roundnose grenadier disturbance responses did not vary significantly between the ROV
transects VT3-1 in the Bay of Biscay and ME16-1 on the Mid Atlantic Ridge (Table 1). In both
cases only few individuals were recorded as being entirely undisturbed (10-21 %) and 67-87 %
of all the disturbed individuals encountered responded to the vehicles at far distance. Both the
locomotion behavior and the vertical positioning registered prior to disturbance responses
differed significantly between the two habitats (locomotion: p=0.0005; vertical positioning:
p<0.025). Roundnose grenadier occurred much higher above the bottom and showed a much
higher rate of drifting on the ridge site. Station holding was frequently registered in the Bay of
Biscay habitat, whereas it did not occur on the ridge site.
c. Differences in disturbance response and natural behavior between habitats and species
(Fig. 6b)
232 Autonomous Underwater Vehicles

The false boarfish showed significantly more disturbance responses (p<0.05), reacting more
frequently at far distances during ROV transect VT1-1 in the Bay of Biscay compared to
ROV transect ME4-2-1 on the Mid-Atlantic Ridge. In the latter habitat this species was
positioned slightly higher above the bottom (p=0.08 for well- and far-above bottom
categories combined) and showed a significant difference in locomotion behavior (p<0.005)
with much less station holding and a higher rate of forward movement. Compared to the co-
occurring codling in the Bay of Biscay transect, false boarfish showed significantly less
disturbance (p<0.005), a much more frequent positioning well or far above the bottom
(p<0.0001) and significantly more drifting and less station holding (p<0.0001)

Codling VT3-2
a 100
DISTURBANCE RESPONSE
b 100
DISTURBANCE RESPONSE

80 No response 80 No response
Close distance Close distance
Far distance
% of Activity

% of Activity
Far distance
Arriving disturbed 60 Arriving disturbed
60

40 40

20 20

0 0
100 100
VERTICAL POSITION VERTICAL POSITION
Close to bottom Close to bottom
80 80
Well above bottom Well above bottom
Far above bottom
% of Activity

Far above bottom

% of Activity

60 60

40 40

20 20
.
0 0
70 100
LOCOMOTION LOCOMOTION
60 Inactive Inactive
Drifting 80 Drifting
50 Station holding Station holding
% of Activity

% of Activity

Forward moving Forward moving

40 60

30
40

20
20
10

0 . 0
ME10-1 ME10-3 Roundnose Codling
grenadier

Fig. 5. Disturbance responses, vertical positioning above bottom and locomotion behavior in
(a) codling during two ROV transects on the northern Mid-Atlantic Ridge and (b)
roundnose grenadier and codling during a ROV transect in Belle Isle Canyon, Bay of Biscay
Deep-Sea Fish Behavioral Responses to Underwater Vehicles:
Differences Among Vehicles, Habitats and Species 233

.
Roundnose grenadier False boarfish Codling
a 100
DISTURBANCE RESPONSE
b 100
DISTURBANCE RESPONSE

80 No response 80 No response
Close distance Close distance

% of Activity
% of Activity

Far distance Far distance

60 Arriving disturbed 60 Arriving disturbed

40 40

20 20

.
0 . 0
100 100
VERTICAL POSITION VERTICAL POSITION

80 Close to bottom 80 Close to bottom

Well above bottom Well above bottom

% of Activity
% of Activity

Far above bottom Far above bottom

60 60

40 40

20 .
20

0 0
70 100
LOCOMOTION LOCOMOTION
60 Inactive Inactive
Drifting 80 Drifting
Station holding Station holding
50
% of Activity

Forward moving
% of Activity

Forward moving
60
40

30
40

20
20
10

0 0
VT3-2 ME16-1 VT1-1 ME4-2-1 VT1-1
Fig. 6. Disturbance responses, vertical positioning above bottom and locomotion behavior in
(a) roundnose grenadier during two ROV transects in Belle Isle Canyon, Bay of Biscay (left),
and on the northern Mid-Atlantic Ridge (right) and (b) false boarfish (left and middle) and
codling (right) during two transects, one on Mériadzek Terrace, Bay of Biscay (VT1-1) and
the other one on the Mid-Atlantic Ridge (ME4-2-1)

4. Discussion and conclusions

Deep-sea fish disturbance responses
The underwater vehicles involved in this study elicited disturbance responses in deep-sea
fishes encountered during bottom transects that can be best interpreted as avoidance or flight
behavior. Clear signs of attraction to the UV’s as they have been reported elsewhere (e.g.,
Stoner et al. 2007; Moore et al. 2008) were not observed. No longer vehicle stops and no point
or selective long-term observations (e.g., by following individual fish) were conducted during
the dive transects. In the studies presented here behavioral recordings were only made during
234 Autonomous Underwater Vehicles

the fishes’ appearance on the forward directing video screen during transects. It is well
possible that additional disturbance responses occurred at larger distances before appearance
or after the fish disappeared on the screen, but those were not recorded. Apart from these
obvious restrictions, the registration and subsequent quantitative comparison of disturbance
responses recorded during UV video transects is a solid method to investigate the influences
of various factors such as different vehicles, habitats, or species on the frequency and intensity
of evoked reactions (see also, Lorance et al. 2002, Uiblein et al. 2002, 2003).
While the manned submersible did not evoke any response in codling (first case study), they
responded considerably disturbed when encountered in the same area with an ROV. A large
portion of the disturbance responses happened at far distance or even before encounter
indicating early detection, before the main illumination focus reached the fish. Sound may
therefore be seen as a main source of disturbance. No exact comparative measurements are
however available of the light and sound intensity produced by the two vehicles during those
dives. Also, the possibilities cannot be ruled out that the signals acted in combination and that
other disturbance sources such as, e.g., pressure waves produced by the moving vehicle body
were involved, too. The present findings provide however no evidence that the much larger-
bodied manned submersible elicited a comparatively higher disturbance response than any of
the four ROV’s used, whereas an opposite effect was demonstrated in the first case study.
In orange roughy, light may play an important role in addition to sound in eliciting
disturbance responses, because a considerable portion of the reactions occurred at short
distances only. Interestingly, the responsiveness to the ROV Bathysaurus decreased between
the two adjacent habitats on the ridge. No differences in natural behavior (vertical positioning
and locomotion) were observed. One additional difference, however, was a much higher
density of orange roughy during the first transect, indicating aggregation formation. Does
orange roughy remain particularly vigilant when residing in dense conspecific aggregations?
During transects with the manned submersible Nautile an aggregation of orange roughy in the
central St. Nazaire canyon did not differ in disturbance responses from conspecifics
encountered in the peripheral area (Lorance et al. 2002, Uiblein et al. 2003). Aggregation
formation in this species may be related to rather different activities such as resting, spawning
or feeding (Lorance et al. 2002). More detailed studies of this ROV dive in the area of the
northern Mid-Atlantic Ridge are planned that shall also include comparisons with roundnose
grenadier and associated habitat conditions encountered during these transects.
Depth may be an important factor influencing disturbance responses, as can be concluded
from the behavior of codling during ROV transects in the Bay of Biscay. These results
corroborate with behavioral observations of the northern cutthroat eel Synaphobranchus
kaupii which also showed more frequent disturbance responses at a deeper located dive in
the Bay of Biscay (Uiblein et al. 2002, 2003). The latter species shows a deeper-bigger pattern,
hence larger fish living at greater depth have a larger sensory surface that should facilitate
signal perception. Also, as food becomes scarcer with larger depths, fish need to pay more
attention to environmental stimuli. Both these argumentations may also apply to codling,
however, more field and biological data would be necessary to test these assumptions.
Species differences in disturbance responses during single dive transects provide the best
evidence for the importance of intrinsic, organism-dependent factors that need to be
considered when studying anthropogenic disturbance. Codling showed no response during
the manned submersible dive in the Bay of Biscay (OB22), while roundnose grenadier
responded considerably and hence may be more sensitive to the signals emitted from this
vehicle. It reacted mainly at far distance or immediately before encounter what points towards
the perception of rather far-ranging signals (e.g., rather noise than light). On the other hand,
Deep-Sea Fish Behavioral Responses to Underwater Vehicles:
Differences Among Vehicles, Habitats and Species 235

codling showed considerable disturbance responses when confronted with an ROV. In the
same situation, false boarfish responded to a lesser extent. These three taxa differ
fundamentally from each other in their biology: the codling typically holds station close to soft
bottoms, the false boarfish prefers to swim or drift closely to shelter provided by hard bottom
structures and corals, while the roundnose grenadier is more flexible showing different
locomotion behavior and vertical positioning depending on habitat context. Among these
three species/species groups, false boarfish appears least prone to predation risk, also given
their rather high body (see also Moore et al. 2008). Probably the response to UV’s reveals also
something about a species’ vigilance and assessment of predation risk.
Deep-sea fish disturbance responses and natural behavior: the full picture
When disturbance responses are properly identified, recorded and analyzed, natural
behaviour can be studied separately thus allowing to gain insights into the ecology of deep-
sea fishes even in the presence of anthropogenic influences. To illustrate this, four case
studies were conducted, three elaborating different aspects of natural behavior (locomotion,
vertical positioning) with disturbance effects remaining constant and one with all three
behaviors varying. In the first two instances only locomotion varied for codling between
two separated transects during an ROV dive on the Mid-Atlantic Ridge and for roundnose
grenadier and codling during a single ROV transect in the Bay of Biscay. These data indicate
that while species clearly differ among each other (“species-specific” behavior), it is also of
high importance to understand their behavioral flexibility in adaptation to different habitats.
Behavioral flexibility or plasticity allows a choice among different locomotion modes and to
select those that fit best to the prevailing conditions in the respective habitat. For instance,
less station holding and increased inactivity (“sit and wait”) as exemplified by codling in
one of two ridge habitats (Fig. 5a) should allow efficient, energy-saving foraging when
currents are weak or absent and food abundance is relatively high.
As deep-sea fishes are behaviorally flexible, one can expect to find considerable differences
among contrasting habitats, as demonstrated for the roundnose grenadier by ROV dives in
the Bay of Biscay and the Mid-Atlantic Ridge. While disturbance responses remained rather
similar in both areas, the fish displayed more drifting and no station holding and were
positioned significantly higher in the water column on the ridge. This reflects obviously
behavioral adjustment to typical ridge conditions (see also, Zaferman 1992) with food
particles arriving at the bottom mainly through the water column, while food input deriving
from the productive shelf areas is lacking.
A rather complex picture of deep-sea fish behavioral ecology is obtained when all behaviors
differ and different habitats are contrasted with different species or species groups, like in
the last case study. False boarfish from habitats in the Bay of Biscay and the Mid-Atlantic
Ridge were compared showing less disturbance responses, a slightly higher vertical
position, less station holding, and more forward movement on the ridge site. The boarfish’s
behavior in the Bay of Biscay clearly contrasts with codling during the same transect, the
latter showing a higher disturbance response, a position on or very close to the bottom, and
more station holding. Interpretations are however complicated through one (or several)
additional factor(s) that need to be considered in this as well as in the anterior case study
featuring roundnose grenadier, because two different UV’s were used.
Towards optimizing in situ behavioral ecology of deep-sea fishes and related research
A promising approach towards reaching best possible interpretations of what deep-sea fishes
do, why they do it, and how they respond to human-induced environmental changes is to
consider all influential external and internal factors in the data analysis and in the
236 Autonomous Underwater Vehicles

interpretation of the results. The central method to approach this goal is to analyze video-
recordings made during UV transects based on detailed description, categorization and
registration of the entire behavior observed with special emphasis to separate human-induced
responses from natural behavior, followed by statistical comparisons. Additional data on the
biology and ecology of the target species, the physical and biological environment, and the
effects and possible impacts of anthropogenic disturbance need to be integrated, too.
To reduce complexity, the number of influential variables should be minimized whenever
possible. Optimally, the same design models of UV’s should be used during all dives that
need to be compared. Dive transects, video recordings, and data analysis should follow
standardized protocols. During each transect representative size measurements combined
with estimates of absolute swimming speed should be obtained from each studied fish
species. Visually well identifiable species should be preferably selected for study so to
minimize possible informational noise introduced by species differences within composed
groups. Groups of closely related species should be used only exceptionally, when in situ
species identification is impossible and the species have a very similar body structure, hence
similar behavior can be expected. Short video or photographic close-ups of each individual
fish from problematic species groups should be taken (preferably by a second camera) to
visualize diagnostic details helpful for species identification. Advice and assistance from
taxonomists specialized in problematic fish groups should be gathered.
Use of different UV’s in comparative studies cannot be recommended, because it may turn
out to be difficult, if not impossible, to adjust for disturbance effects. Most certainly more
than a single signal source of disturbance needs to be considered. Experimental
manipulation of light, sound, and vehicle velocity, but possibly also the magnetic field,
singly or in combination, might certainly assist to better understand the relative importance
of these potential sources of disturbance (see also Stoner et al. 2007). However, for full
control of disturbance effects from UV’s, one would also need to investigate the receiver bias
and in particular the sensory equipment (Popper & Hastings 2009) and reaction norms
(Tuomainen & Candolin 2010) which may differ considerably among fish species,
populations, size classes, and ontogenetic stages.
The longer the encounter with an UV the more increases the chance of interactions and
evocation of disturbance responses. During longer UV stops, odor plumes deriving from
collected organisms or bait brought along may be formed and scavengers may be attracted
(Trenkel & Lorance 2011). If point observation are made during longer stops of an UV, these
data should be treated separately from transect data. Also, when stationary, the vehicle itself
may be perceived in quite different ways than when transitionally encountered during
transects and disturbance responses may change and in some cases shift from avoidance to
attraction or even to aggression (see for instance, Moore et al. 2008). Observations of deep-sea
fishes deriving from longer-term interactions with UV’s are certainly interesting per se, but
may not always contribute to properly understand natural behavior. To reduce interactions it
may be of advantage to position the vehicle firmly on the ground and switch off the motors for
behavioral observations close to the bottom. During point observations in the open water as
well as close to the bottom switching off the illumination and use of infrared light combined
with infrared-sensitive cameras should be considered (Widder et al. 2005).
As stated initially, investigations of the effects of UV’s on deep-sea fish behavior have
important implications for many other studies of deep-sea fishes, as for instance, in situ
assessments of abundances, populations dynamics, habitat associations, community
structure, and patterns of biological diversity (Stoner et al. 2008). Hence the suggestions and
Deep-Sea Fish Behavioral Responses to Underwater Vehicles:
Differences Among Vehicles, Habitats and Species 237

recommendations towards optimization of in situ behavioral ecology may prove useful also
for broader applications in deep-sea fish research and management.

5. Summary
An important prerequisite for in situ ecological investigations of deep-sea fishes using
underwater vehicles (UV’s) is to distinguish between disturbance responses elicited by the
vehicles and undisturbed natural behavior. Nine case studies deriving from ten video transects
along deep bottoms of the North Atlantic (Bay of Biscay, Mid-Atlantic Ridge) with a manned
submersible and three remotely operated vehicles (ROV’s) are presented to demonstrate
differences in behavioral disturbance between vehicles, habitats, and species. Three species,
roundnose grenadier (Coryphaenoides rupestris), orange roughy (Hoplostethus atlanticus) and
false boarfish (Neocyttus helgae), and codling, a group of closely related species (North Atlantic
codling, Lepidion eques, being the most common), were studied. During each UV transect
recordings of disturbance responses and two activity patterns shown by undisturbed fishes,
vertical positioning in the water column and locomotion mode, were made. Each behavior was
subdivided into several categories and analyzed quantitatively using sample sizes larger than
18 individuals per species/species group and transect. Codling showed no disturbance
responses to a manned submersible, while reacting intensely to a ROV during two transects
performed in the same area. When the same UV was used, clear differences in disturbance
responses were found between both adjacent dive transects and species/species groups
indicating habitat- and species-specific responsiveness to signals emitted by the vehicle, in
particular sound and light, but possibly also other sources. In three additional case studies,
disturbance responses remained rather constant between transects or species, but natural
behavior differed. The final study provides the fullest picture with all three behaviors
differing, the interpretations being however complicated by the fact that different vehicles
were used in different habitats. The findings are discussed emphasizing the significance of in
situ quantitative behavioral studies of UV-based video transects in deep-sea fish ecology and
related research fields. Detailed suggestions and recommendations towards optimization of
vehicle-disturbance control and observation techniques are provided.

6. Acknowledgements
Travel support provided by a bilateral Amadeus project (between Austria and France,) no.
V13, made comparative studies of video footage from the Bay of Biscay UV dives possible.
Special thanks to Daniel Latrouite, Pascal Lorance and Verena Trenkel, IFREMER, who
provided copies of relevant video footage from OBSERVHAL and VITAL dives and
assistance in behavioral data recordings. Thanks to Jan Bryn, Reidar Johannesen, and Asgeir
Steinsland for assistance during MAR-ECO ROV dives. Thanks to Mirjam Bachler for
valuable comments on the manuscript.

7. References
Beale, C.M. 2007. The Behavioral Ecology of Disturbance Responses. International Journal of
Comparative Psychology 20: 111-120
Beebe, W. 1933. Preliminary Account of Deep Sea Dives in the Bathysphere with Especial
Reference to One of 2200 Feet. Proceedings of the National Academy of Sciences of
the USA. 1933 January; 19(1): 178–188.
238 Autonomous Underwater Vehicles

Heezen, B.C. & C.D. Hollister. 1971. The Face of the Deep. Oxford UP, New York.
Lorance, P., Trenkel, V.M. & F. Uiblein, 2005. Characterizing natural and reaction behaviour
of large mid-slope species: consequences for catchability. In: Shotton, R. (Ed). 2005.
Deep Sea 2003: Conference on the Governance and Management of Deep-sea
Fisheries. Part1: Conference Reports. FAO Fisheries Proceedings 3/1: 162-164
Lorance, P. & V. Trenkel, 2006. Variability in natural behaviour, and observed reactions to
an ROV, by mid-slope fish species. Journal of Experimental Marine Biology and
Ecology 332 : 106-119.
Lorance, P., Uiblein, F., & D. Latrouite, 2002: Habitat, behaviour and colour patterns of
orange roughy Hoplostethus atlanticus (Pisces: Trachichthyidae) in the Bay of Biscay.
Journal of the Marine Biological Association of the UK 82: 321-331
Moore, J.A., Auster, P.J., Calini, D., Heinonen, K., Barber, K. & B. Hecker, 2008. False
boarfish Neocyttus helgae in the Western North Atlantic. Bulletin of the Peabody
Museum of Natural History 49: 31-41.
Popper, A.N. & M.C. Hastings 2009; The effects of human-generated sound on fish.
Integrative Zoology 4: 43-52
Sokal, R.R. & F.J. Rohlf, 1981. Biometry. WH Freeman, New York.
Stoner, A.W., Ryer, C.H., Parker, S.J., Auster, P.J. & W.W. Wakefield, 2008. Evaluating the
role of fish behavior in surveys conducted with underwater vehicles. Canadian
Journal of Fisheries and Aquatic Sciences 65: 1230-1241
Trenkel V.M., Francis, R.I.C.C., Lorance, P. Mahévas, S., Rochet M-J. & D.M. Tracey. 2004a.
Availability of deep-water fish to trawling and visual observation from a remotely
operated vehicle (ROV). Marine Ecology Progress Series 284:293–303.
Trenkel V.M., Lorance P. & S. Mahévas 2004b. Do visual transects provide true population
density estimates for deep-water fish? ICES Journal of Marine Science 62:1050–1056.
Trenkel V.M., Lorance P., 2011. Estimating Synaphobranchus kaupii densities; Contribution of
fish behavior to differences between bait experiments and visual strip transects.
Deep-Sea Research I 58:63-71.
Tuomainen, U. & U. Candolin, 2011. Behavioural responses to human-induced
environmental change. Biological Reviews, 86:640-657
Uiblein, F., Lorance, P. & D. Latrouite, 2002: Variation in locomotion behaviour in northern
cutthroat eel (Synaphobranchus kaupi) on the Bay of Biscay continental slope. Deep-
Sea Research I 49: 1689-1703
Uiblein, F., Lorance, P. & D. Latrouite 2003: Behaviour and habitat utilization of seven
demersal fish species on the Bay of Biscay continental slope, NE Atlantic. Marine
Ecology Progress Series 257: 223–232
Uiblein, F., Bordes, F., Lorance, P., Nielsen, J.G., Shale, D., Youngbluth, M. & R.
Wienerroither, 2010: Behavior and habitat selection of deep-sea fishes: a
methodological perspective. In: S. Uchida (ed.): Proceedings of an International
Symposium Into the Unknown, researching mysterious deep-sea animals 2007,
Okinawa Churaumi Aquarium, Okinawa, Japan, 5-21.
Widder, E.A., Robison, B.H., Reisenbichler, K.R. & S.H.D. Haddock, 2005. Using red light for
in situ observations of deep-sea fishes. Deep-Sea Research I 52: 2077–2085.
Yoklavich, M.M., Love, M.S. & K.A. Forney, 2007. A fishery-independent assessment of an
overfished rockfish stock, cowcod (Sebastes levis), using direct observations from an
occupied submersible. Canadian Journal of Fisheries and Aquatic Sciences 64: 1795-
1804.
 11

Mapping and Dilution Estimation of Wastewater

Discharges Based on Geostatistics Using an
Autonomous Underwater Vehicle
Patrícia Ramos1 and Nuno Abreu2
1 Institutefor Systems and Computer Engineering of Porto (INESC Porto)
School of Accounting and Administration of Porto – Polytechnic Institute of Porto
2 Institute for Systems and Computer Engineering of Porto (INESC Porto)

Faculty of Engineering of University of Porto

Portugal

1. Introduction
Wastewaters are often discharged into coastal waters through outfall diffusers that efficiently
dilute effluent and usually restrict any environmental impact within a small area. However,
predicting this impact is difficult because of the complexity of the hydrodynamic processes
that mix the wastewater and also because of the variability in oceanic conditions. Despite great
improvements over the years in the understanding of these mixing processes, since models are
now available that can make reasonable predictions under steady-state conditions (Hunt et al.,
2010), many aspects remain unknown and unpredictable. For this reason, much effort has
been recently devoted to improve ways of monitoring and characterizing sewage plumes
under a variety of oceanographic conditions.

1.1 MARES AUV

Autonomous Underwater Vehicles (AUVs) have been used efficiently in a wide range of
applications. They were first developed with military applications in mind, for example
for mine hunting missions. Later on, scientists realized their true potential and started
to use them as mobile sensors, taking measurements in difficult scenarios and at a
reasonable cost (Bellingham, 1997; Bellingham et al., 1992; Fernandes et al., 2000; Nadis, 1997;
Robinson et al., 1999; Yu et al., 2002). MARES (Modular Autonomous Robot for Environment
Sampling) AUV has been successfully used to monitor sea outfalls discharges (Abreu et al.,
2010; Abreu & Ramos, 2010; Ramos & Abreu, 2010; 2011a;b;c) (see Fig. 1). MARES is 1.5 m
long, has a diameter of 8-inch and weighs about 40 kg in air. It features a plastic hull with
a dry mid body (for electronics and batteries) and additional rings to accommodate sensors
and actuators. Its modular structure simplifies the system’s development (the case of adding
sensors, for example). It is propelled by two horizontal thrusters located at the rear and two
vertical thrusters, one at the front and the other at the rear. This configuration allows for small
operational speeds and high maneuverability, including pure vertical motions. It is equipped
with an omnidirectional acoustic transducer and an electronic system that allows for long
baseline navigation. The vehicle can be programmed to follow predefined trajectories while
240
2 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

collecting relevant data using the onboard sensors. A Sea-Bird Electronics 49 FastCAT CTD
had already been installed onboard the MARES AUV to measure conductivity, temperature
and depth. MARES’ missions for environmental monitoring of wastewater discharges are
conducted using a GUI software that fully automates the operational procedures of the
campaign (Abreu et al., 2010). By providing visual and audio information, this software
guides the user through a series of steps which include: (1) real time data acquisition from
CTD and ADCP sensors, (2) effluent plume parameter modeling using the CTD and ADCP
data collected, (3) automatic path creation using the plume model parameters, (4) acoustic
buoys and vehicle deployment, (5) automatic acoustic network setup and (6) real time tracking
of the AUV mission.

Fig. 1. Autonomous Underwater Vehicle MARES.

1.2 Data processing

Data processing is the last step of a sewage outfall discharge monitoring campaign. This
processing involves the ability to extrapolate from monitoring samples to unsampled
locations. Although very chaotic due to turbulent diffusion, the effluent’s dispersion process
tends to a natural variability mode when the plume stops rising and the intensity of turbulent
fluctuations approaches to zero (Hunt et al., 2010). It is likely that after this point the pollutant
substances are spatially correlated. In this case, geostatistics appears to be an appropriate
technique to model the spatial distribution of the effluent. In fact, geostatistics has been
used with success to analyze and characterize the spatial variability of soil properties, to
obtain information for assessing water and wind resources, to design sampling strategies
for monitoring estuarine sediments, to study the thickness of effluent-affected sediment in
the vicinity of wastewater discharges, to obtain information about the spatial distribution
of sewage pollution in coastal sediments, among others. As well as giving the estimated
values, geostatistics provides a measure of the accuracy of the estimate in the form of the
kriging variance. This is one of the advantages of geostatistics over traditional methods
of assessing pollution. In this work ordinary block kriging is used to model and map the
spatial distribution of temperature and salinity measurements gathered by an AUV on a
Portuguese sea outfall monitoring campaign. The aim is to distinguish the effluent plume
from the receiving waters, characterize its spatial variability in the vicinity of the discharge
and estimate dilution.
Mapping and Dilution Estimation of Wastewater Discharges
Based on
Mapping and Geostatistics
Dilution Using
Estimation of Wastewater an Autonomous
Discharges Underwater
based on Geostatistics Vehicle
using an Autonomous Underwater Vehicle 2413

2. Geostatistical analysis
2.1 Stationary random function models
The most widely used geostatistical estimation procedures use stationary random function
models. A random function is a set of random variables that have some spatial locations and
whose dependence on each other is specified by some probabilistic mechanism. A random
function is stationary if all the random variables have the same probability distribution
and if any pair of random variables has a joint probability distribution that depends only
on the separation between the two points and not on their locations. If the random
function is stationary, then the expected value and the variance can be used to summarize
the univariate behavior of the set of random variables. The parameter that is commonly
used to summarize the bivariate behavior of a stationary random function is its covariance
function, its correlogram, and its variogram. The complete definition of the probabilistic
generating mechanism of a random function is usually difficult even in one dimension.
Fortunately, for many of the problems we typically encounter, we do not need to know the
probabilistic generating mechanism. We usually adopt a stationary random function as our
model and specify only its covariance or variogram (Isaaks & Srivastava, 1989; Kitanidis, 1997;
Wackernagel, 2003).

2.2 Ordinary kriging

Ordinary kriging method is often referred with the acronym BLUE which stands for “Best
Linear Unbiased Estimator”. “Linear” because its estimates are weighted linear combinations
of the available data; “Unbiased” since it tries to have the mean error equal to 0; and
“Best”because it aims at minimizing the variance of the errors. Let us then see how the
concept of a random function model can be used to decide how to weight the nearby
samples so that our estimates are unbiased. For any point at which we want to estimate
the unknown value, our model is a stationary random function that consists of n random
variables, one for the value at each of the n sample locations, Z (x1 ), Z (x2 ), . . . , Z (xn ), and one
for the unknown value at the point we are trying to estimate Z (x0 ). Each of these random
variables has the same probability law; at all locations, the expected value of the random
variable is m and the variance is σ2 . Every value in this model is seen as an outcome (or
realization) of the random variable. Our estimate is also a random variable since it is a
weighted linear combination of the random variables at the n sampled locations (Cressie,
1993; Goovaerts, 1997; Isaaks & Srivastava, 1989; Kitanidis, 1997; Stein, 1999; Wackernagel,
2003; Webster & Oliver, 2007):
n
Ẑ(x0 ) = ∑ w i · Z ( x i ). (1)
i =1

The estimation error is defined as the difference between the random variable modeling the
true value and the estimate:
ε(x0 ) = Z (x0 ) − Ẑ (x0 ). (2)

The estimation error is also a random variable. Its expected value, often referred to as the bias,
is  
n
E [ ε(x0 )] = m 1− ∑ wi . (3)
i =1
242
4 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

Setting this expected value to 0, to ensure an unbiasedness estimate results in:

n
∑ wi = 1. (4)
i =1

This is known as the condition of unbiasedness (Isaaks & Srivastava, 1989; Kitanidis, 1997;
Wackernagel, 2003). The expression of the variance of the modeled error is
   
σε2( x0 ) = var [ Z (x0 )] − 2 cov Ẑ (x0 ), Z (x0 ) + var Ẑ(x0 ) . (5)

Since we have already assumed that all of the random variables have the same variance σ2 ,
then var [ Z (x0 )] = σ2 . The second term in Equation 5 can be written as

  n
− 2 cov Ẑ (x0 ), Z (x0 ) = −2 ∑ wi · cov [ Z (xi ), Z (x0 )]
i =1
n
= −2 ∑ wi Ci0. (6)
i =1

The third term of Equation 5 can be expressed as

  n n  
var Ẑ (x0 ) = ∑ ∑ wi · w j · cov Z ( x i ), Z ( x j )
i =1 j =1
n n
= ∑ ∑ wi w j Cij . (7)
i =1 j =1

Then, the expression of the error variance comes in the following way:
n n n
σε2( x0 ) = σ2 − 2 ∑ wi Ci0 + ∑ ∑ wi w j Cij . (8)
i =1 i =1 j =1

Equation 8 expresses the error variance as function of the n weights, once chosen the
random model function parameters, namely the variance σ2 and all the covariances Cij . The
minimization of σε2( x ) is constrained by the unbiasedness condition imposed earlier, which can
0
be solved using the method of Lagrange multipliers. We start by introducing a new parameter
μ, called the Lagrange multiplier, in Equation 8 in the following way:
 
n n n n
σε2( x0 ) = σ2 − 2 ∑ wi Ci0 + ∑ ∑ wi w j Cij + 2 μ ∑ wi − 1 . (9)
i =1 i =1 j =1 i =1

=0

Then we minimize σε2( x ) by calculating the n + 1 partial first derivatives of Equation 9 with
0
respect to the n weights and the Lagrange multiplier, and setting each one to 0, which
Mapping and Dilution Estimation of Wastewater Discharges
Based on
Mapping and Geostatistics
Dilution Using
Estimation of Wastewater an Autonomous
Discharges Underwater
based on Geostatistics Vehicle
using an Autonomous Underwater Vehicle 2435

produces the following system of equations:

∂(σε2( x ) ) n n

∂ ( w1 )
0
= −2 C10 + 2 ∑ w j C1j + 2 μ = 0 ⇒ ∑ w j C1j + μ = C10
j =1 j =1
.. ..
. .
∂(σε2( x ) ) n n

∂(wi )
0
= −2 Ci0 + 2 ∑ w j Cij + 2 μ = 0 ⇒ ∑ w j Cij + μ = Ci0
j =1 j =1
.. ..
. .
∂(σε2( x ) ) n n

∂(wn )
0
= −2 Cn0 + 2 ∑ w j Cnj + 2 μ = 0 ⇒ ∑ w j Cnj + μ = Cn0
j =1 j =1
 
∂(σε2( x ) ) n n

∂(μ )
0
=2 ∑ wi − 1 =0 ⇒ ∑ wi = 1
i =1 i =1

which can also be written in a compact way as

n n
∑ w j Cij + μ = Ci0, ∀i = 1, 2, . . . , n; ∑ wi = 1. (10)
j =1 i =1

This system of equations, often referred to as the ordinary kriging system, can be written in
matrix notation as
C · W = D
⎡ ⎤ ⎡ ⎤⎤ ⎡
C11 . . . C1n 1 w1 C10
⎢ .. .. . . .. ⎥ ⎢ .. ⎥ ⎢ .. ⎥
⎢ . . . .⎥ ⎢ ⎥ ⎢ ⎥
⎢ ⎥· ⎢ . ⎥ =⎢ . ⎥ (11)
⎣ C . . . Cnn 1 ⎦ ⎣ wn ⎦ ⎣C ⎦
n1 n0
1 ... 1 0 μ 1
  
( n+1)×( n+1) ( n+1)×1 ( n+1)×1
The set of weights and the Lagrange multiplier that will produce an unbiased estimate of
Z (x0 ) with the minimum error variance are then given by

W = C−1 · D. (12)

The value of σε2( x ) can be obtained in a quicker way using an alternative expression to Eq. 8.
0
Multiplying each of the n equations given in Eq. 10 by wi and summing these n equations
leads to the following:
n n n
∑ ∑ wi w j Cij = ∑ wi Ci0 − μ.
i =1 j =1 i =1

Substituting this into Equation 8 the minimized error variance comes as follows:
 
n
σε2( x0 ) = σ2 − ∑ wi Ci0 + μ , (13)
i =1
244
6 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

or, in terms of matrices as

σε2( x0 ) = σ2 − W T · D. (14)
The minimized error variance is usually called the ordinary kriging variance.

2.3 Block kriging

A consideration in many environmental applications has been that ordinary kriging usually
exhibits large prediction errors (Bivand et al., 2008). This is due to the larger variability in
the observations. When predicting averages over larger areas, i.e. within blocks, much of the
variability averages out and consequently block mean values have lower prediction errors. If
the blocks are not too large the spatial patterns do not disappear. The block kriging system
is similar to the point kriging system given by Equation 11. The matrix C is the same since
it is independent of the location at which the block estimate is required. The covariances
for the vector D are point-to-block covariances. Supposing that the mean value over a block
V is approximated by the arithmetic average of the N point variables contained within that
block (Goovaerts, 1997; Isaaks & Srivastava, 1989), i.e.

1 N
N j∑
ZV ≈ Z ( x j ), (15)
=1

the point-to-block covariances required for vector D are

1 N
N j∑
CiV = cov [ Z (xi ), ZV ] = Cij , ∀i = 1, 2, . . . , n. (16)
=1

The block kriging variance is

 
n
σV2 = CVV − ∑ wi CiV + μ , (17)
i =1

where CVV is the average covariance between pairs of points within V:

1 N N
N 2 i∑ ∑ Cij .
CVV = (18)
=1 j =1

An equivalent procedure, that can be computationally more expensive than block kriging,
is to obtain the block estimate by averaging the N kriged point estimates within the
block (Goovaerts, 1997; Isaaks & Srivastava, 1989).

2.4 Spatial continuity

Spatial continuity exists in most earth science data sets. When we look at a contour map,
or anything similar, the values do not appear to be randomly located, but rather, low
values tend to be near other low values and high values tend to be near other high values.
I.e. two measurements close to each other are most likely to have similar values than
two measurements far apart (Isaaks & Srivastava, 1989). To compute the set of weights
and the Lagrange multiplier, that will produce each estimate and the resulting minimized
error variance, we need to know the covariances of C and D matrices. As we said before,
since our random function is stationary, all pairs of random variables separated by a
Mapping and Dilution Estimation of Wastewater Discharges
Based on
Mapping and Geostatistics
Dilution Using
Estimation of Wastewater an Autonomous
Discharges Underwater
based on Geostatistics Vehicle
using an Autonomous Underwater Vehicle 2457

distance and direction h (known as lag) have the same joint probability distribution. The
covariance function, C (h ) is the covariance between random variables separated by a lag
h (Isaaks & Srivastava, 1989; Kitanidis, 1997; Wackernagel, 2003). For a stationary random
function, the covariance function C (h ) is:

C (h ) = E [ Z (x) Z (x + h )] − {E [ Z (x)]}2 . (19)

The covariance between random variables at identical locations is the variance of the random
function:  
C (0) = E { Z (x)}2 − {E [ Z (x)]}2 = var [ Z (x)] = σ2 . (20)
The semivariogram, or simply variogram, is half the expected squared difference between
random variables separated by a lag h:

1   1
γ (h ) = E { Z (x) − Z (x + h )}2 = var [ Z (x) − Z (x + h )] . (21)
2 2
The quantity γ (h ) is known as the semivariance at lag h. The “semi”refers to the fact that it
is half of a variance. The variogram between random variables at identical locations is zero,
i.e. γ (0) = 0. Using Equations 19, 20 and 21, we can relate the variogram with the covariance
function as:
γ ( h ) = C ( 0 ) − C ( h ) = σ2 − C ( h ). (22)
In practice, the pattern of spatial continuity chosen for the random function is usually taken
from the spatial continuity evident in the sample data set. Geostatisticians usually define
the spatial continuity of the sample data set through the variogram and solve the ordinary
kriging system using covariance (Isaaks & Srivastava, 1989). The maximum value reached by
the variogram is called the sill. The distance at which the sill is reached is called the range. The
vertical jump from zero at the origin to the value of semivariance at extremely small separation
distances is called the nugget effect. The estimator of the variogram usually used, known
as Matheron’s method-of-moments estimator (MME) is (Matheron, 1965; Webster & Oliver,
2007)
N (h)
1
2N (h ) i∑
[ Z (xi ) − Z (xi + h )]2 ,
γ (h ) = (23)
=1
where z(xi ) is the value of the variable of interest at location xi and N (h ) is the number of pairs
of points separated by the particular lag vector h. Cressie and Hawkins (Cressie & Hawkins,
1980) developed an estimator of the variogram that should be robust to the presence of
outliers and enhance the variogram spatial continuity, having also the advantage of not
spreading the effect of outliers in computing the maps. This estimator (CRE) is defined as
follows (Cressie & Hawkins, 1980):
 4
N (h)
1
N (h ) i∑
| Z (xi ) − Z (xi + h )|1/2
1 =1
γ (h ) = × . (24)
2 0.457 + N 0.494
(h)
+ 0.045 2
[ N ( h )]

Once the sample variogram has been calculated, a function (called the variogram model) has
to be fit to it. First, because the matrices C and D may need semivariance values for lags that
are not available from the sample data. And second, because the use of the variogram does
246
8 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

not guarantee the existence and uniqueness of solution to the ordinary kriging system. The
most commonly used variogram models are the spherical model, the exponential model, the
Gaussian model and the Matern model (Isaaks & Srivastava, 1989).

2.5 Cross-validation
Cross-validation is a procedure used to compare the performance of several competing
models (Webster & Oliver, 2007). It starts by splitting the data set into two sets: a modelling
set and a validation set. Then the modelling set is used for variogram modelling and kriging
on the locations of the validation set. Finally the measurements of the validation set are
compared to their predictions (Bivand et al., 2008). If the average of the cross-validation errors
(or Mean Error, ME) is close to 0,

1 m  
m i∑
ME = Z (xi ) − Ẑ (xi ) . (25)
=1

we may say that apparently the estimates are unbiased (Z (xi ) and Ẑ (xi ) are, respectively, the
measurement and estimate at point xi and m is the number of measurements of the validation
set). A significant negative (positive) mean error can represent systematic overestimation
(underestimation). The magnitude of the Root Mean Squared Error (RMSE) is particularly
interesting for comparing different models (Wackernagel, 2003; Webster & Oliver, 2007):

1 m  2
m i∑
RMSE = Z (xi ) − Ẑ(xi ) . (26)
=1

The RMSE value should be as small as possible indicating that estimates are close to
measurements. The kriging standard deviation represents the error predicted by the
estimation method. Dividing the cross-validation error by the corresponding kriging
standard deviation allows to compare the magnitudes of both actual and predicted
error (Wackernagel, 2003; Webster & Oliver, 2007). Therefore, the average of the standardized
squared cross-validation errors (or Mean Standardized Squared Error, MSSE)
 2
1 m Z (xi ) − Ẑ (xi )
m i∑
MSSE = , (27)
=1 σR2 ( x )
i

should be about one, indicating that the model is accurate. A scatterplot of true versus
predicted values provides additional evidence on how well an estimation method has
performed. The coefficient of determination R2 is a good index for summarizing how close
the points on the scatterplot come to falling on the 45-degree line passing through the
origin (Isaaks & Srivastava, 1989). R2 should be close to one.

3. Results
3.1 Study site
A map of the study site is shown in Fig. 2(a). Foz do Arelho outfall is located off the
Portuguese west coast near Óbidos lagoon. In operation since June 2005, is presently
discharging about 0.11 m3 /s of mainly domestic wastewater from the WWTPs of Óbidos,
Carregal, Caldas da Rainha, Gaeiras, Charneca and Foz do Arelho, but it can discharge up to
0.35 m3 /s. The total length of the outfall, including the diffuser, is 2150 m. The outfall pipe,
Mapping and Dilution Estimation of Wastewater Discharges
Based on
Mapping and Geostatistics
Dilution Using
Estimation of Wastewater an Autonomous
Discharges Underwater
based on Geostatistics Vehicle
using an Autonomous Underwater Vehicle 2479

made of HDPE, has a diameter of 710 mm. The diffuser, which consists of 10 ports spaced 8
or 12 meters apart, is 93.5 m long. The ports, nominally 0.175 m in diameter, are discharging
upwards at an angle of 90◦ to the pipe horizontal axis; the port height is about 1 m. The
outfall direction is southeast-northwest (315.5◦ true bearing) and is discharging at a depth of
about 31 m. In that area the coastline itself runs at about a 225◦ angle with respect to true
north and the isobaths are oriented parallel to the coastline. A seawater quality monitoring
program for the outfall has already started in May 2006. Its main purposes are to evaluate
the background seawater quality both in offshore and nearshore locations around the vicinity
of the sea outfall and to follow the impacts of wastewater discharge in the area. During the
campaign the discharge remained fairly constant with an average flowrate of approximately
0.11 m3 /s. The operation area specification was based on the outputs of a plume prediction
model (Hunt et al., 2010) which include mixing zone length, spreading width, maximum rise
height and thickness. The model inputs are, besides the diffuser physical characteristics, the
water column stratification, the current velocity and direction, and the discharge flowrate.
Information on density stratification was obtained from a vertical profile of temperature and
salinity acquired in the vicinity of the diffuser two weeks before the campaign (see Fig. 3).
The water column was weakly stratified due to both low-temperature and salinity variations.
The total difference in density over the water column was about 0.13 σ-unit. The current
direction of 110◦ was estimated based on predictions of wind speed and direction of the day
of the campaign. A current velocity of 0.12 m/s was estimated based on historic data. The
effluent flowrate consider for the plume behavior simulation was 0.11 m3 /s. According to the
predictions of the model, the plume was spreading 1 m from the surface, detached from the
bottom and forming a two-layer flow. The end of the mixing zone length was predicted to
be 141 m downstream from the diffuser. Fig. 2(b) shows the diffuser and a plan view of the
AUV operation area (specified according to the model predictions), mainly in the northeast
direction from the diffuser, covering about 20000 m2 .
The vehicle collected CTD data at 1.5 m and 3 m depth, in accordance to the plume minimum
dilution height prediction. During the mission transited at a fairly constant velocity of 1 m/s
(2 knots) recording data at a rate of 16 Hz. Maximum vertical oscillations of the AUV in
performing the horizontal trajectories were less than 0.5 m (up and down).
100
50
0
North (m)

−50
−100
−150

−50 0 50 100 150

East (m)

(a) Map of the study site (©2011 Google - (b) AUV operation area.
Images).

Fig. 2. Vicinity of Foz do Arelho sea outfall.

248
10 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

Fig. 3. Vertical STD profile used in the plume behavior simulation.

3.2 Exploratory analysis

In order to obtain elementary knowledge about the temperature and salinity data sets,
conventional statistical analysis was conducted (see the results in Table 1 and Table 2). At
the depth of 1.5 m the temperature ranged from 15.359ºC to 15.562ºC and at the depth of
3 m the temperature ranged from 15.393ºC to 15.536ºC. The mean value of the data sets was
15.463ºC and 15.469ºC, respectively at the depths of 1.5 m and 3 m, which was very close to
the median value that was respectively 15.466ºC and 15.472ºC. The coefficient of skewness is
relatively low (-0.309) for the 1.5 m data set and not very high (-0.696) for the 3 m data set,
indicating that in the first case the histogram is approximately symmetric and in the second
case that distribution is only slightly asymmetric. The very low values of the coefficient of
variation (0.002 and 0.001) reflect the fact that the histograms do not have a tail of high values.
At the depth of 1.5 m the salinity ranged from 35.957 psu to 36.003 psu and at the depth of
3 m the salinity ranged from 35.973 psu to 36.008 psu. The mean value of the data sets was
35.991 psu and 35.996 psu, respectively at the depths of 1.5 m and 3 m, which was very close
to the median value that was respectively 35.990 and 35.998 psu. The coefficient of skewness
is not to much high in both data sets (-0.63 and -1.1) indicating that distributions are only
slightly asymmetric. The very low values of the coefficient of variation (0.0002 and 0.0001)
reflect the fact that the histograms do not have a tail of high values. The ordinary kriging
method works better if the distribution of the data values is close to a normal distribution.
Therefore, it is interesting to see how close the distribution of the data values comes to being
normal. Fig. 4 shows the plots of the normal distribution adjusted to the histograms of the
temperature measured at depths of 1.5 m and 3 m, and Fig. 5 shows the plots of the normal
distribution adjusted to the histograms of the salinity measured at depths of 1.5 m and 3 m.
The density value in the histogram is the ratio between the number of samples in a bin and the
total number of samples divided by the width of the bin (constant). Apart from some erratic
high values it can be seen that the histograms are reasonably close to the normal distribution.
Mapping and Dilution Estimation of Wastewater Discharges
Based on
Mapping and Geostatistics
Dilution UsingDischarges
Estimation of Wastewater an Autonomous Underwater
based on Geostatistics Vehicle
using an Autonomous Underwater Vehicle 249
11

[email protected] m [email protected] m
Samples 20,026 10,506
Mean 15.463ºC 15.469ºC
Median 15.466ºC 15.472ºC
Minimum 15.359ºC 15.393ºC
Maximum 15.562ºC 15.536ºC
Coefficient of skewness -0.31 -0.70
Coefficient of variation 0.002 0.001

Table 1. Summary statistics of temperature measurements.

[email protected] m [email protected] m
Samples 20,026 10,506
Mean 35.991 psu 35.996 psu
Median 35.990 psu 35.998 psu
Minimum 35.957 psu 35.973 psu
Maximum 36.003 psu 36.008 psu
Coefficient of skewness -0.63 -1.1
Coefficient of variation 0.0002 0.0001

Table 2. Summary statistics of salinity measurements.

35

35
30

30
25

25
20

20
Density

Density
15

15
10

10
5

5
0

15.35 15.40 15.45 15.50 15.55 15.60 15.35 15.40 15.45 15.50 15.55 15.60

Temperature (°C) Temperature (°C)

Fig. 4. Histograms of temperature measurements at depths of 1.5 m (left) and 3 m (right).

250
12 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

100

100
80

80
60

60
Density

Density
40

40
20

20
0

0
35.95 35.96 35.97 35.98 35.99 36.00 36.01 35.95 35.96 35.97 35.98 35.99 36.00 36.01

Salinity (psu) Salinity (psu)

Fig. 5. Histograms of salinity measurements at depths of 1.5 m (left) and 3 m (right).

3.3 Variogram modeling

For the purpose of this analysis, the temperature and the salinity measurements were divided
into a modeling set (comprising 90% of the samples) and a validation set (comprising 10%
of the samples). Modeling and validation sets were then compared, using Student’s-t test,
to check that they provided unbiased sub-sets of the original data. Furthermore, sample
variograms for the modeling sets were constructed using the MME estimator and the CRE
estimator. This robust estimator was chosen to deal with outliers and enhance the variogram’s
spatial continuity. An estimation of semivariance was carried out using a lag distance of
2 m. Table 3 and Table 4 show the parameters of the fitted models to the omnidirectional
sample variograms constructed using MME and CRE estimators. All the variograms were
fitted to Matern models (for several shape parameters ν) with the exception to the salinity data
measured at the depth of 3 m. The range value (in meters) is an indicator of extension where
autocorrelation exists. The variograms of salinity show significant differences in range. The
autocorrelation distances are always larger for the CRE estimator which may demonstrate the
enhancement of the variogram’s spatial continuity. All variograms have low nugget values
which indicates that local variations could be captured due to the high sampling rate and
to the fact that the variables under study have strong spatial dependence. Anisotropy was
investigated by calculating directional variograms. However, no anisotropy effect could be
shown.

3.4 Cross-validation
The block kriging method was preferred since it produced smaller prediction errors and
smoother maps than the point kriging. Using the 90% modeling sets of the two depths, a
two-dimensional ordinary block kriging, with blocks of 10 × 10 m2 , was applied to estimate
temperature at the locations of the 10% validation sets. The validation results for both
parameters measured at depths of 1.5 m and 3 m depths are shown in Table 5 and Table 6.
At both depths temperature was best estimated by the variogram constructed using CRE.
Salinity at the depth of 1.5 m was best estimated by the variogram constructed using CRE
and at the depth of 3 m was best estimated using the Gaussian model with the MME. The
Mapping and Dilution Estimation of Wastewater Discharges
Based on
Mapping and Geostatistics
Dilution UsingDischarges
Estimation of Wastewater an Autonomous Underwater
based on Geostatistics Vehicle
using an Autonomous Underwater Vehicle 251
13

Variogram
Depth Estimator Model Nugget Sill Range

MME Matern (ν = 0.4) 0.000 0.001 75.0

1.5
CRE Matern (ν = 0.5) 0.000 0.002 80.1
MME Matern (ν = 0.3) 0.000 0.0002 101.3
3.0
CRE Matern (ν = 0.7) 0.000 0.002 107.5

Table 3. Parameters of the fitted variogram models for temperature measured at depths of 1.5
and 3.0 m.

Variogram
Depth Estimator Model Nugget Sill Range

MME Matern (ν = 0.6) 0.436 11.945 134.6

1.5
CRE Matern (ν = 0.6) 0.153 10786.109 51677.1
MME Matern (ν = 0.8) 0.338 11.724 181.6
3.0
CRE Gaussian 0.096 120.578 390.1

Table 4. Parameters of the fitted variogram models for salinity measured at depths of 1.5 and
3 m.

Depth Method R2 ME MSE RMSE

MBK 0.9184 2.0174e-4 8.0530e-5 8.9739e-3
1.5
CBKa 0.9211 1.6758e-4 7.7880e-5 8.8248e-3
MBK 0.8748 1.0338e-4 3.6295e-5 6.0244e-3
3.0
CBKa 0.8827 0.6538e-4 3.4008e-5 5.8316e-3
a The preferred model.
Table 5. Cross-validation results for the temperature maps at depths of 1.5 and 3 m.

difference in performance between the two estimators: block kriging using the MME estimator
(MBK) or block kriging using the CRE estimator (CBK) is not substantial. Fig. 6 shows the
omnidirectional sample variograms for temperature at the depth of 1.5 m and 3 m fitted by
the preferred models. Fig. 7 shows the omnidirectional sample variograms for salinity at the
depth of 1.5 m and 3 m fitted by the preferred models.
Fig. 8 and Fig. 9 show the scatterplots of true versus estimated values for the most satisfactory
models. The dark line is the 45º line passing through the origin and the discontinuous line
is the OLS (Ordinary Least Squares) regression line. These plots show that observed and
predicted values are highly positively correlated. The R2 value for the temperature at the
depth of 1.5 m was 0.9211 and the RMSE was 0.0088248ºC, and at the depth of 3 m was 0.8827
and the RMSE was 0.0058316ºC (Table 5). The R2 value for the salinity at the depth of 1.5 m
was 0.9513 and the RMSE was 0.0016435 psu, and at the depth of 3 m was 0.8982 and the
RMSE was 0.0019793 psu (Table 6).
252
14 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

Depth Method R2 ME MSE RMSE

MBK 0.9471 3.1113e-5 2.8721e-6 1.6947e-3
1.5
CBKa 0.9513 -3.1579e-5 2.7010e-6 1.6435e-3
MBKa 0.8982 -7.1735e-5 3.9175e-6 1.9793e-3
3.0
CBK 0.7853 -8.1264e-5 8.2589e-6 2.8738e-3
a The preferred model.
Table 6. Cross-validation results for the salinity maps at depths of 1.5 and 3 m.

0.0008

0.0010
Semivariance (°C2)

Semivariance (°C2)

0.0006

0.0004

0.0005

0.0002

20 40 60 80 100 120 20 40 60 80 100 120

Distance (m) Distance (m)

Fig. 6. Variograms for temperature at depths of 1.5 m (left) and 3 m (right).

12

10
6

8
Semivariance (psu2)

Semivariance (psu2)

4
6

20 40 60 80 100 120 20 40 60 80 100 120

Distance (m) Distance (m)

Fig. 7. Variograms for salinity at depths of 1.5 m (left) and 3 m (right).

Mapping and Dilution Estimation of Wastewater Discharges
Based on
Mapping and Geostatistics
Dilution UsingDischarges
Estimation of Wastewater an Autonomous Underwater
based on Geostatistics Vehicle
using an Autonomous Underwater Vehicle 253
15

15.60

15.60
15.55

15.55
Predicted temperature (°C)

Predicted temperature (°C)

15.50

15.50
15.45

15.45
15.40

15.40
15.35

15.35
15.35 15.40 15.45 15.50 15.55 15.60 15.35 15.40 15.45 15.50 15.55 15.60

Observed temperature (°C) Observed temperature (°C)

Fig. 8. Predicted versus observed temperature at the depths of 1.5 m (left) and 3 m (right)
using the preferred models.
36.01

36.01
36.00

36.00
35.99

35.99
Predicted salinity (psu)

Predicted salinity (psu)

35.98

35.98
35.97

35.97
35.96

35.96
35.95

35.95

35.95 35.96 35.97 35.98 35.99 36.00 36.01 35.95 35.96 35.97 35.98 35.99 36.00 36.01

Observed salinity (psu) Observed salinity (psu)

Fig. 9. Predicted versus observed salinity at the depths of 1.5 m (left) and 3 m (right) using
the preferred models.

3.5 Mapping
Fig. 10 shows the block kriged maps of temperature on a 2 × 2 m2 grid using the preferred
models. Fig. 13 shows the block kriged maps of salinity on a 2 × 2 m2 grid using the preferred
models. In the 1.5 m kriged map the temperature ranges between 15.407ºC and 15.523ºC and
the average value is 15.469ºC (the measured range is 15.359ºC–15.562ºC and the average value
is 15.463ºC). In the 3 m kriged map the temperature ranges between 15.429ºC and 15.502ºC
and the average value is 15.467ºC (the measured range is 15.393ºC–15.536ºC and the average
value is 15.469ºC). We may say that estimated values are in accordance with the measurements
since their distributions are similar (identical average values, medians, and quartiles). The
difference in the ranges width is due to only 5.0% of the samples in the 1.5 m depth map
(2.5% on each side of the distribution) and only 5.3% of the samples in the 3.0 m depth map
254
16 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

(3.1% on the left side and 2.2% on the rigth side of the distribution). These samples should
then be identified as outliers not representing the behaviour of the plume in the established
area. In the 1.5 m kriged map the salinity ranges between 35.960 psu and 36.004 psu and the
average value is 35.992 psu, which is in accordance with the measurements (the measured
range is 35.957psu – 36.003psu and the average value is 35.991 psu). In the 3 m kriged map
the salinity ranges between 35.977 psu and 36.004 psu and the average value is 35.995 psu,
which is in accordance with the measurements (the measured range is 35.973psu – 36.008psu
and the average value is 35.996 psu). As predicted by the plume prediction model, the effluent
was found dispersing close to the surface. From the temperature and salinity kriged maps it is
possible to distinguish the effluent plume from the background waters. It appears as a region
of lower temperature and lower salinity when compared to the surrounding ocean waters
at the same depth. At the depth of 1.5 m the major difference in temperature compared to
the surrounding waters is about -0.116ºC while at the depth of 3 m this difference is about
-0.073ºC. At the depth of 1.5 m the major difference in salinity compared to the surrounding
waters is about -0.044 psu while at the depth of 3 m this difference is about -0.027 psu. It is
important to note that these very small differences in temperature and salinity were detected
due to the high resolution of the CTD sensor. (Washburn et al., 1992) observed temperature
and salinity anomalies in the plume in the order, respectively of -0.3ºC and -0.1 psu, when
compared with the surrounding waters within the same depth range. The small plume-related
anomalies observed in the maps are evidence of the rapid mixing process. Due to the
large differences in density between the rising effluent plume and ambient ocean waters,
entrainment and mixing processes are vigorous and the properties within the plume change
rapidly (Petrenko et al., 1998; Washburn et al., 1992). The effluent plume was found northeast
from the diffuser beginning, spreading downstream in the direction of current. Using the
navigation data, we could later estimate current velocity and direction and the values found
were, respectively, 0.4 m/s and 70ºC, which is in accordance with the location of the plume.

15.52 15.52

50 15.50 50 15.50

15.48 15.48

0 0
North (m)

North (m)

15.46 15.46

−50 −50
15.44 15.44

15.42 15.42
−100 −100

15.40 15.40

−150 −150
15.38 15.38
0 50 100 150 0 50 100 150

East (m) East (m)

Fig. 10. Prediction map of temperature distribution (ºC) at depths of 1.5 m (left) and 3 m
(right).
Fig. 12 shows the variance of the estimation error (kriging variance) for the maps of
temperature distribution at depths of 1.5 m and 3 m. The standard deviation of the estimation
error is less than 0.0195ºC at the depth of 1.5 m and less than 0.0111ºC at the depth of 3
Mapping and Dilution Estimation of Wastewater Discharges
Based on
Mapping and Geostatistics
Dilution UsingDischarges
Estimation of Wastewater an Autonomous Underwater
based on Geostatistics Vehicle
using an Autonomous Underwater Vehicle 255
17

36.00 36.00

50 50

35.99 35.99

0 0

35.98 35.98
North (m)

North (m)
−50 −50

35.97 35.97

−100 −100
35.96 35.96

−150 −150
35.95 35.95
0 50 100 150 0 50 100 150

East (m) East (m)

Fig. 11. Prediction map of salinity distribution (psu) at depths of 1.5 m (left) and 3 m (right).

m. Results of the same order were obtained for salinity. It’s interesting to observe that,
as expected, the variance of the estimation error is less the closer is the prediction from the
trajectory of the vehicle. The dark blue regions correspond to the trajectory of MARES AUV.

0.00035 0.00035

50 50
0.00030 0.00030

0.00025 0.00025
0 0
North (m)

North (m)

0.00020 0.00020

−50 −50
0.00015 0.00015

0.00010 0.00010
−100 −100

0.00005 0.00005

−150 −150
0.00000 0.00000
0 50 100 150 0 50 100 150

East (m) East (m)

Fig. 12. Variance of the estimation error for the maps of temperature distribution at depths of
1.5 m (left) and 3 m (right)

3.6 Dilution estimation

Environmental effects are all related to concentration C of a particular contaminant X.
Defining Ca as the background concentration of substance X in ambient water and C0 as the
concentration of X in the effluent discharge, the local dilution comes as follows (Fischer et al.,
1979):
C − Ca
S= 0 , (28)
C − Ca
256
18 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

   
which can be rearranged to give C = Ca S− S
1 + 1 C . In the case of variability of the
S 0
background concentration of substance X in ambient water the local dilution is given by

C0 − Ca0
S= , (29)
C − Ca
where Ca0 is the background concentration of substance X in ambient
 water at the discharge
depth. This expression in 29 can be arranged to give C = Ca + S1 (C0 − Ca0 ), which in
simple terms means that the increment of concentration above background is reduced by the
dilution factor S from the point of discharge to the point of measurement of C. Using salinity
distribution at depths of 1.5 m and 3 m we estimated dilution using Equation 29 (see the
contour maps in Fig. 13). We assumed C0 = 2.3 psu, Ca0 = 35.93 psu, Ca = 36.008 psu at
1.5 m depth and Ca = 36.006 psu at 3 m depth. The minimum dilution estimated at the depth
of 1.5 m was 705 and at the depth of 3.0 m was 1164 which is in accordance with Portuguese
legislation that suggests that outfalls should be designed to assure a minimum dilution of
50 when the plume reaches surface (INAG, 1998). (Since dilution increases with the plume
rising we should expect that the minimum values would be greater if the plume reached
surface (Hunt et al., 2010)).

8000 16000

50 7000 50 14000

6000 12000

0 0

5000 10000
North (m)

North (m)

4000 8000
−50 −50

3000 6000

−100 −100
2000 4000

1000 2000
−150 −150

0 50 100 150 0 50 100 150

East (m) East (m)

Fig. 13. Dilution maps at depths of 1.5 m (left) and 3 m (right).

4. Conclusion
Through geostatistical analysis of temperature and salinity obtained by an AUV at depths
of 1.5 m and 3 m in an ocean outfall monitoring campaign it was possible to produce kriged
maps of the sewage dispersion in the field. The spatial variability of the sampled data has been
analyzed and the results indicated an approximated normal distribution of the temperature
and salinity measurements, which is desirable. The Matheron’s classical estimator and Cressie
and Hawkins’ robust estimator were then used to compute the omnidirectional variograms
that were fitted to Matern models (for several shape parameters) and to a Gaussian model.
The performance of each competing model was compared using a split-sample approach.
In the case of temperature, the validation results, using a two-dimensional ordinary block
Mapping and Dilution Estimation of Wastewater Discharges
Based on
Mapping and Geostatistics
Dilution UsingDischarges
Estimation of Wastewater an Autonomous Underwater
based on Geostatistics Vehicle
using an Autonomous Underwater Vehicle 257
19

kriging, suggested the Matern model (ν = 0.5 − 1.5 m and ν = 0.7 − 3.0 m) with semivariance
estimated by CRE. In the case of salinity, the validation results, using a two-dimensional
ordinary block kriging, suggested the Matern model (ν = 0.6 − 1.5 m and ν = 0.8 − 3.0 m)
with semivariance estimated by CRE, for the depth of 1.5 m, and with semivariance estimated
by MME, for the depth of 3 m. The difference in performance between the two estimators
was not substantial. Block kriged maps of temperature and salinity at depths of 1.5 m and
3 m show the spatial variation of these parameters in the area studied and from them it is
possible to identify the effluent plume that appears as a region of lower temperature and lower
salinity when compared to the surrounding waters, northeast from the diffuser beginning,
spreading downstream in the direction of current. Using salinity distribution at depths of 1.5
m and 3 m we estimated dilution at those depths. The values found are in accordance with
Portuguese legislation. The results presented demonstrate that geostatistical methodology
can provide good estimates of the dispersion of effluent that are very valuable in assessing the
environmental impact and managing sea outfalls.

5. Acknowledgment
This work was partially funded by the Foundation for Science and Technology (FCT)
under the Program for Research Projects in all scientific areas (Programa de Projectos
de Investigação em todos os domínos científicos) in the context of WWECO project -
Environmental Assessment and Modeling of Wastewater Discharges using Autonomous
Underwater Vehicles Bio-optical Observations (Ref. PTDC/MAR/74059/2006).

6. References
Abreu, N., Matos, A., Ramos, P. & Cruz, N. (2010). Automatic interface for AUV mission
planning and supervision, MTS/IEEE International Conference Oceans 2010, Seattle,
USA.
Abreu, N. & Ramos, P. (2010). An integrated application for geostatistical analysis of sea
outfall discharges based on R software, MTS/IEEE International Conference Oceans
2010, Seattle, USA.
Bellingham, J. (1997). New Oceanographic Uses of Autonomous Underwater Vehicles, Marine
Technology Society Journal 31(3): 34–47.
Bellingham, J., Goudey, C., Consi, T. R. & Chryssostomidis, C. (1992). A Small Long Range
Vehicle for Deep Ocean Exploration, Proceedings of the International Offshore and Polar
Engineering Conference, San Francisco, pp. 151–159.
Bivand, R. S., Pebesma, E. J. & Gómez-Rubio, V. (2008). Applied spatial data analysis with R.
ISBN: 97 -0-387-78170-9.
Cressie, N. (1993). Statistics for spatial data, New York.
Cressie, N. & Hawkins, D. M. (1980). Robust estimation of the variogram, I, Jour. Int. Assoc.
Math. Geol. 12(2): 115–125.
Fernandes, P. G., Brierley, A. S., Simmonds, E. J., Millard, N. W., McPhail, S. D., Armstrong,
F., Stevenson, P. & Squires, M. (2000). Fish do not Avoid Survey Vessels, Nature
404: 35–36.
Fischer, H. B., List, J. E., Koh, R. C. Y., Imberger, J. & Brooks, N. H. (1979). Mixing in Inland and
Coastal Waters, Academic Press.
Goovaerts, P. (1997). Geostatistics for natural resources evaluation, Applied Geostatistics Series.
ISBN13: 9780195115383, ISBN10: 0195115384.
258
20 Autonomous Underwater Vehicles
Will-be-set-by-IN-TECH

Hunt, C. D., Mansfield, A. D., Mickelson, M. J., Albro, C. S., Geyer, W. R. & Roberts, P.
J. W. (2010). Plume tracking and dilution of effluent from the Boston sewage outfall,
Marine Environmental Research 70: 150–161.
INAG (1998). Linhas de Orientação Metodológia para a Elaboração de Estudos Técnicos Necessários
para Cumprir o Art� 7� do D.L. 152/97. Descargas em Zonas Menos Sensíveis,
Instituto da Água. Ministério do Ambiente, Lisboa.
Isaaks, E. H. & Srivastava, R. M. (1989). Applied geostatistics, New York Oxford. ISBN
0-19-505012-6-ISBN 0-19-505013-4 (pbk.).
Kitanidis, P. (1997). Introduction to geostatistics: applications in hydrogeology, New York (USA).
Matheron, G. (1965). Les variables régionalisées et leur estimation: une application de la théorie des
fonctions aléatoires aux sciences de la nature, Paris.
Nadis, S. (1997). Real Time Oceanography Adapts to Sea Changes, Science 275: 1881–1882.
Petrenko, A. A., Jones, B. H. & Dickey, T. D. (1998). Shape and Initial Dilution of Sand Island,
Hawaii Sewage Plume, Journal of Hydraulic Engineering, ASCE 124(6): 565–571.
Ramos, P. & Abreu, N. (2010). Spatial analysis of sea outfall discharges using block kriging,
6th International Conference on Marine Waste Water Discharges and Coastal Environment,
MWWD2010, Langkawi, Malaysia.
Ramos, P. & Abreu, N. (2011a). Environmental impact assessment of Foz do Arelho sewage
plume using MARES AUV, IEEE/OES International Conference Oceans 2011, Santander,
Spain.
Ramos, P. & Abreu, N. (2011b). Spatial analysis and dilution estimation of Foz do Arelho
Outfall using observations gathered by an AUV, International Symposium on Outfall
Systems, Mar del Plata, Argentina.
Ramos, P. & Abreu, N. (2011c). Using an AUV for Assessing Wastewater Discharges Impact:
An Approach Based on Geostatistics, Marine Technology Society Journal 45(2).
Robinson, A., Bellingham, J., Chryssostomidis, C., Dickey, T., Levine, E., Petrikalakis, N.,
Porter, D., Rothschild, B., Schmidt, H., Sherman, K., Holliday, D. & Atwood, D.
(1999). Real-time forecasting of the multidisciplinary coastal ocean with the littoral
ocean observing and predicting system (LOOPS), Proceedings of the Third Conference on
Coastal Atmospheric and Oceanic Prediction Processes, American Meteorological Society,
New Orleans, LA.
Stein, M. L. (1999). Interpolation of spatial data: some theory for kriging, New York.
Wackernagel, H. (2003). Multivariate geostatistics: an introduction with applications, Berlin.
Washburn, L., Jones, B. H., Bratkovich, A., Dickey, T. D. & Chen, M.-S. (1992). Mixing,
Dispersion, and Resuspension in Vicinity of Ocean Wastewater Plume, Journal of
Hydraulic Engineering, ASCE 118(1): 38–58.
Webster, R. & Oliver, M. (2007). Geostatistics for environmental scientists, 2nd Edition. ISBN-13:
978-0-470-02858-2(HB).
Yu, X., Dickey, T., Bellingham, J., Manov, D. & Streitlien, K. (2002). The Application
of Autonomous Underwater Vehicles for Interdisciplinary Measurements in
Massachusetts and Cape Cod Bayes, Continental Shelf Research 22(15): 2225–2245.