Cui et al.

Intelligent Marine
Intelligent Marine Technology and Systems (2023) 1:7
https://doi.org/10.1007/s44295-023-00010-3 Technology and Systems

REVIEW Open Access

Review of research and control technology

of underwater bionic robots
Zhongao Cui1†, Liao Li1†, Yuhang Wang1, Zhiwei Zhong1 and Junyang Li1*

Abstract
As marine resources continue to be exploited, the remarkable locomotion and coordination of fish provide an excel-
lent source of inspiration for scientists and engineers to design and control the next -generation autonomous under-
water vehicles within a bionic framework. Underwater biomimetic robots combine bionics and robot technology,
and their biological characteristics offer a lot of convenience for the robot so that it can obtain better performance
in adaptability and robustness. Recently, with the combination of bionics, mechanics, electronics, materials science,
and automation, there has been great progress in developing underwater bionic robots with different structure types
and energy supply modes. This paper summarizes the research status of underwater robots, focuses on the research
status of underwater bionic robots with different materials, types and motion modes, and introduces the propulsion
mechanism of underwater robots with different structures and the control methods adopted in the propulsion pro-
cess. Finally, the broad application prospect and market potential of underwater biomimetic robot are introduced.
Keywords Underwater bionic robot, Propulsion mechanism, Control strategies, Application scenarios

1 Introduction remotely operated vehicles (ROVs), and cableless under-

A bionic underwater robot, as the name suggests, is a water robots, traditionally known as autonomous under-
new type of robot that imitates the propulsion mecha- water vehicles (AUVs) (Wynn et al. 2014). Moreover, they
nism and body structure of fish or other marine creatures can be categorized by use into underwater investigation
living underwater using electromechanical components robots (observation, measurement, test material collec-
and intelligent materials (such as memory alloy materials, tion, etc.) and underwater operation robots (underwater
mixed materials, and rigid materials), which can adapt to welding, pipe twisting, underwater construction, under-
different underwater environments and realize underwa- water cutting, etc.) (Vu et al. 2018).
ter propulsion (Chu et al. 2012). It has the characteristics At present, most underwater robots are frame-based,
of high efficiency, high mobility, and low noise (Chen similar to the rotating elongated body of a submarine.
et al. 2021a). For a long time, scholars have been com- With the continuous development of bionic technology,
mitted to studying marine biological propulsion models the bionic fish shape and control modes of underwater
and bionic underwater robots. Underwater vehicles can robots will also evolve (Xie et al. 2021). In this review,
be classified into two groups based on their structural different control system algorithms are described, such
design: cabled underwater vehicles, commonly known as as those developed for individual or cluster control of
underwater robots (Khalaji and Zahedifar 2020). Under-
†
water robots work in unknown and challenging marine
Zhongao Cui and Liao Li contributed equally to this work.
environments. Complex marine environments, such as
*Correspondence: wind, waves, currents, and water pressure, severely inter-
Junyang Li
junyangli3-c@my.cityu.edu.hk fere with robot motion and control, making communica-
1
College of Information Science and Engineering, Ocean University tion and navigation of underwater robots very difficult.
of China, Qingdao 266100, China Thus, the development potential of underwater robots

© The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://creativecommons.org/licenses/by/4.0/.
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 2 of 28

still needs to be continuously explored. In this paper, generate enough thrust to propel itself in static water
various underwater robots are reviewed and introduced conditions (Villanueva et al. 2011) (Fig. 1a). Harbin
from manufacturing materials, structural design, drive Engineering University developed an underwater jel-
mechanism, and control strategy. lyfish microrobot prototype model based on SMA and
ionic conductive polymer film (ICPF) as actuators to
1.1 Shape memory alloy material achieve swimming movement, with an overall size
Shape memory alloys (SMAs) are solid, smart materi- of microrobot of about 75 mm long, 55 mm in diam-
als driven by current silently. The principle of operation eter, and 6.5 g weight. This tiny jellyfish-like robot has
is that when heated from low-temperature martensite four tentacles (Fig. 1b). Each mechanism consists of a
to high-temperature austenite, SMA returns to the pre- restraint mechanism and an ICPF actuator, and each
determined shape and generates activation, a process tentacle can work with an SMA driver to increase its
known as the reversed-phase transition (Yang et al. range of motion and provide greater propulsion. The
2023). When cooled from austenite to martensite, SMA energized SMA shrinks the internal volume of the
experiences a martensitic phase transition and returns to microrobot so that the water or other water-contain-
its initial state under bias stress (Hu et al. 2023). Previ- ing medium inside the microrobot is driven backward,
ous works reported that the bionic starfish robot and the thus forming a propulsion force (Yang et al. 2007).
bionic manta ray robot fish were both powered by SMA. Another jet-propelled jellyfish bionic robot, MPA-O,
The jellyfish robot, or Robojelly, developed by Vir- developed by Harbin Engineering University, is made
ginia Tech, is driven by a bionic shape memory alloy of SMA material. The length of the moving direction is
composite actuator modeled after a jellyfish. With a 46.1 mm, and the section diameter is 36.3 mm. At an
body made of RTV silicone with a total mass of 242 g operating frequency of 0.6 Hz, the robot has a maxi-
and a bell-shaped diameter of 164 mm, Robojelly can mum speed of 6 mm/s (Guo et al. 2007) (Fig. 1c).

Fig. 1 SMA material underwater bionic robot. a Jellyfish robot (Villanueva et al. 2011); b Jellyfish-like microrobot (Yang et al. 2007); c Jellyfish-like
micro robot that achieves MPA-O swimming mode using a hybrid actuator (Guo et al. 2007); d Structure of the jellyfish-like robot (Ko et al. 2012);
e Microrobot manta ray (Wang et al. 2009); f Fish skeleton structure including a latex-based skin for water protection (Rossi et al. 2011)
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 3 of 28

The miniature jellyfish swimming robot, powered by multifunctional characteristics of natural organisms.
the SMA system developed by Chonnam National Uni- Flexible robots comprising SMA materials can achieve
versity, has four flexible fins, each equipped with a per- elastic deformation and pass through narrow spaces
manent magnet for electromagnetic drive, and the body without causing internal damage (Hu et al. 2019).
of the robot has a length of 17 mm and a thickness of
0.5 mm. The SMA driver can generate a uniform mag- 1.2 Ionic polymer–metal composite
netic field in the desired direction in 3D space, which Biomimetic artificial muscle material is a new type of
can bend the fins of the jellyfish-like microrobot. Thus, intelligent material developed rapidly in the 1990s, con-
the cyclic changes in the uniform magnetic field will stantly setting off a global research upsurge and has
synchronize the fluctuations of the fins and generate important application value in the aerospace, biomimetic
a propulsive force for the robot in the desired direc- robot, and biomedical engineering fields. Ionic polymer-
tion (Ko et al. 2012) (Fig. 1d). A miniature bionic manta metal composites (IPMC), as electrochemical actuators,
ray robotic fish with a triangular pectoral fin driven by are typical biomimetic artificial muscle materials (Şafak
SMA, developed by the Harbin Institute of Technology, and Adams 2002) with a sandwich structure comprising
is based on the simplified pectoral fin model described two layers of electrodes and ionic polymers. Under the
in the joint. Each consists of a bionic fin on the leading electric field, electrical and mechanical energy can be
edge and a latex film (0.2 mm thick) that forms the sur- converted by the reversible dissociation process of ions
face of the fin. The front part of the tail is attached to the at the electrode interface. IPMC material has the advan-
body and is a flexible fin that can adjust the course (Wang tages of fast response, large drive displacement, and low
et al. 2009) (Fig. 1e). The SMA is used as a continuous drive voltage. However, it can only be used in wet envi-
backbone for curved fish, and the University of Madrid ronments and has huge limited applications in amphibi-
utilized six SMA-based actuators to make the skeleton of ous bionic robots (Cao et al. 2022).
the robotic fish. Their length is 1/3 of their body length IPMC is widely used in the manufacture of body or
(8.5 cm, excluding the tail fins and head). They are posi- caudal-fin (BDF) swimming robots, and the fish-like
tioned in pairs parallel to the body so that antagonistic robot developed by the New York Institute of Technology
motion can enable the robotic fish to resist higher water is designed with IPMC materials, which mimics the gen-
pressure (Rossi et al. 2011) (Fig. 1f ). eral swimming movement of a fish, protecting itself with
Jellyfish robots developed at Kagawa University its tail fin (Marras and Porfiri 2012). In 2010, Michigan
achieved flutter-like motion with SMA-based actuators State University developed a wireless bionic robotic fish
and positive spring elements (Najem et al. 2012). The that also successfully demonstrated the swimming mode
robot was 63 mm long, 35 mm wide, and 18 mm high, the of the BCF using a robot based on IPMC as the driving
same size as the manta ray robot previously developed by material, which had four different types of fins mounted
the laboratory, which was constructed with SMA wires on its tail to optimize the relationship between robot
embedded in an elastic substrate (Xie et al. 2018). The jel- speed and fin shape (Brown and Clark 2010). The Uni-
lyfish robot’s fluctuating motion is generated by 10 SMA versity of Science and Technology of China employed the
actuators (5 on each side) with a maximum speed of IPMC brake as the tail fin of the robot fish for propul-
40 mm/s at 3.125 Hz, and the Department of Health Sci- sion, which mainly comprises two servo motors, namely,
ences and Technology of ETH Zurich (Vogel 2012) devel- angle rotating block pairs and brakes, and the main body
oped another jellyfish-like robot with jet propulsion via is arranged symmetrically. The experiment confirms that
an SMA-based actuator. The drive is called bionic shape the robot fish with the two degrees of freedom caudal-fin
memory alloy composite (BISMAC). BISMAC is assem- propulsion mechanism can realize various basic swim-
bled from a steel spring and SMA wire embedded in a ming movements by using a caudal fin (Zhou et al. 2017)
silicone precision connection (Ma et al. 2019) and has a (Fig. 2a). The University of Nevada studied a biomimetic
length of 110 mm and thickness of 0.1 mm, while the bell jellyfish robot powered by an IPMC placed inside a sili-
has diameters of 134 and 82 mm. The robot is 242 mm cone dome. Because the selected jellyfish shell elastomer
long, 225 mm wide, and 52 mm high. Because its main material is soft enough, the IMP can be easily driven
component is silicone, it has greater hardness, strong without hindrance, and the robot can increase thrust
compressive performance, and a speed of 35 mm/s. production by approximately 1300% compared to a nor-
Traditional robots are machined from rigid materials, mal jellyfish (Trabia et al. 2016) (Fig. 2b).
which often limits their ability to deform and adapt their The University of Virginia developed a biomi-
shape to the external environment. Although these rigid metic robot that mimics a manta ray’s pectoral fin,
robots have the advantages of large output force, high which partially comprises a PDMS (polydimethylsi-
precision, and strong controllability, they often lack the loxane) membrane using four IPMCs (0.28 mm thick)
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 4 of 28

Fig. 2 IPMC material robot fish. a Prototype of the robotic fish (Zhou et al. 2017); b Robot jellyfish prototype (Trabia et al. 2016); c IPMC material
jellyfish robot (Yeom and Oh 2008); d Robot fish propelled by IPMC (Hu and Zhou 2009)

(Sankaranarayanan et al. 2008). Jeonnam National Uni- applications. In the past decade, research and application
versity utilized IPMC actuators to build and evaluate of robotic artificial muscles have been developed (Wynn
biomimetic jellyfish robots. The existing IPMC actuators et al. 2014). More fundamental research is required
limit their application fields due to their flat shape, which regarding how artificial muscles can be manufactured,
is a severe defect of this actuator material. To overcome modeled, controlled, and engineered to acquire fish-like
the disadvantages of planar IPMC actuators, a curved muscle properties and achieve muscle-like behavior.
IPMC actuator with predetermined initial deformation is
developed. The expected initial deformation is acquired 1.3 Piezoelectric composite ceramics
by heat treatment. The bionic input signal is generated by A group of researchers from the Artificial Muscle
imitating the real movement of jellyfish (Yeom and Oh Research Center at Konkuk University in South Korea
2008) (Fig. 2c). The experiment confirms that the jellyfish developed a bionic fish robot that utilizes its tail fin to
robot can move normally. IPMC is an important electro- drive the swimming motion of BCF (Pham et al. 2023)
active polymer (artificial muscle) with built-in drive and and used a lightweight piezoelectric composite ceramic
sensing capabilities. The robotic fish developed by the (PZT), a single crystal piezoelectric ceramic encased
Intelligent Microsystems Laboratory at Michigan State in glass/epoxy and carbon/epoxy resins, as the actuator
University comprises IMPC material. Its motion mecha- material. The robot uses a crank, rack, and pinion struc-
nism includes the following: IPMC usually includes a thin ture (the size of the robot is increased by 400 mm due
ion exchange membrane, which is chemically plated on to the need for an additional device to achieve the move-
two surfaces with precious metals as electrodes. Appli- ment). The robot has a maximum speed of 25.16 mm/s
cation of voltage to the IPMC leads to the transport of at an operating voltage of 300 vpp and an operating fre-
hydrated cations and water molecules within the mem- quency of 0.9 Hz.
brane and the associated electrostatic interactions to The miniature underwater mobile robot developed by
result in bending motion, leading to a driving effect (Hu Professor Toshio Fukuda of Nagoya University in Japan
and Zhou 2009) (Fig. 2d). The artificial muscles of biomi- contains piezoelectric ceramics to drive the oscillation
metic robots’ balance drive the performance, power-to- of two symmetrical legs to realize its movement. The two
weight ratio, and muscle form factors. As such, they are legs of the robot are equipped with a pair of symmetri-
ideally suited as biomimetic actuators for various robotic cal fins at a certain angle. The symmetrical structural
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 5 of 28

design can offset the lateral force and strengthen the for- and high in efficiency. This kind of robot has broad appli-
ward momentum. A 250-fold elastic hinge amplification cation prospects and value in microtubule detection and
mechanism is designed to amplify the PZT. The robot is biomedicine.
320 mm long and 190 mm wide, and the motion speed
is 21.6–32.5 m/s. In 2009, the Marine Science Center of 1.4 Mixed materials
Northeastern University in the United States developed The underwater environment is complex, so the mate-
a robot fish with wave propulsion using PZT materi- rial requirements of the underwater bionic robot are
als and chain rod structure (Zhou et al. 2008). Through very strict. Currently, polymer-metal composite mate-
lateral body fluctuations, the robotic eel drives itself rials are widely employed, combining the advantages
through the water column and controls its floating depth. of both the polymer and the metal. Polymers can with-
In 2008, DRAPER Lab launched VCUUV (Vorticity Con- stand a certain degree of deformation in most envi-
trol Unmanned Underwater Vehicle), a piezoelectric ronments. Both materials can make good adjustments
ceramic-driven robot fish designed after tuna (Suk and to the impact of the external environment, with the
Hwan 2014). It is about 2.4 m long and weighs 300 lbs. Its polymer being lighter and the metal material being
maximum swing frequency is 1.5 Hz, with a maximum harder (Zheng et al. 2020). The robot fish developed
swimming speed of 1.25 m/s at 1 Hz. The goal of the by Shandong University of Technology is made of a
laboratory is to develop an autonomous underwater vehi- resin-mixed material and a rigid motor, with four main
cle using eddy current-controlled propulsion and show parts: two laminated tail fins, a rigid fish body with a
that PZT materials have good drag reduction character- permanent magnet at the tail, a miniature battery,
istics, excellent maneuverability, depth-holding ability, and a controller. During the driving process, electri-
and higher acceleration and deceleration ability through cal energy is converted into mechanical energy of the
free swimming. Harbin Engineering University (Yue et al. tail fin, producing the swimming motion of the robotic
2015) developed a water microrobot with a PZT drive; fish (Yan et al. 2021) (Fig. 3a). Kagawa University has
the main feature of this drive is that it is a polymer mate- developed a medusa-like underwater bionic micro-
rial actuator only in water or wet environment work. The robot based on SMA and artificial muscles. It moves
robot, which can be turned forward, left, or right, has a like a jellyfish, floats and sinks, and has two pecto-
pair of driving wings driven by a pulse voltage to gener- ral fins to achieve swimming motion (Shi et al. 2010)
ate propulsion. PZT material can realize the continuous (Fig. 3b). The mollusk developed by Zhejiang Univer-
fluctuation deformation of the fluctuating fins of bionic sity includes a steering electronic server, a steering tail,
fish, which makes it compact in structure, light in weight, and two SMA flapping wings. Two dielectric elastomer

Fig. 3 Mixed-material robot. a Composite robotic fish structure (Yan et al. 2021); b Prototype jellyfish-like biomimetic underwater microrobot (Shi
et al. 2010); c Mechanism composition of the soft robotic fish (Zhang et al. 2021); d Dolphin robot (Shen et al. 2013); e Composite robotic fish (Xie
et al. 2020); f Illustration of the robotic fish (Marras and Porfiri 2012)
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 6 of 28

(DE) membranes are clamped onto the electrodes to driven by the servo motor generate the periodic sinusoi-
form an artificial muscle. Precut frames and precut dal movement of the flexible polyester tail fin to mimic
rebar are glued to each side of the muscle. The pur- the movement of fish (Marras and Porfiri 2012) (Fig. 3f ).
pose of the insulation board is to prevent the feed pipe
to make contact with the support frame. The flexible 2 Underwater robot control system classification
wavy fins provide power when the wings are in the flap- At present, the commonly used motion control methods
ping, stretching, or actuating state. When AC voltage is of underwater biomimetic robots are model-based con-
applied, the wing changes back and forth between the trol methods, sine controllers, and central mode genera-
former state and the driven state, providing forward tor (CPG)-based methods. As the structural components
force (Zhang et al. 2021) (Fig. 3c). of marine biomimetic robots usually include power
DE, which is widely used in robot drives, has good modules, sensors, chips, and driving components, the
softness, and its outstanding advantages are that the behavior of the bionic robot is controlled by the prede-
relative adjustment rate after shape change is fast, the fined program or the command controller (the power
response is quite rapid, the energy consumption is less, supply of the controller is mainly provided by traditional
and the mechanical and electrical conversion efficien- lithium batteries) (Chen et al. 2021b). Depth adjustment
cies are high. The dielectric elastic material-driven robot of the robot in water is controlled by the controller and
developed by Kagawa University is a jet propulsion robot is mainly completed by the buoyancy unit. Thus, con-
simulating a pike (Bal et al. 2019). The driver is com- trol can also be divided into rigid motor control and soft
posed of SMA, ICPF, and rubber materials. The length drive control.
of the motion direction is 46.1 mm, the diameter of
the section is 36.3 mm, and the maximum speed of the 2.1 Model control method
robot is 6 mm/s. The dolphin robot developed by Bei- The model control method is performed by analyzing the
hang University consists of three parts: (1) a rigid plas- dynamics and kinematics of the robot and then establish-
tic shell that acts as a body, (2) IPMC stripes that act as ing a complex mathematical model. The mathematical
muscles, and (3) a plastic sheet that mimics a tail fin. The model can accurately calculate the next movement of the
shell is designed based on the proportions of the dol- robot to achieve the effect of precise control. However,
phin’s streamlined body, made of nylon plastic, using a due to the complex and changeable underwater envi-
3D printer, and covered with a black matte resin varnish, ronment, accurately modeling the robot is very difficult.
leading to a smooth surface. The IPMC is attached to the Even if it can be accurately modeled, its control mode
body by two small rectangular conductive copper plates, is very complex. In 2014, Inner Mongolia University of
which act as clamps, with a flexible fin attached to the Technology (Li et al. 2014b) developed a set of integrated
end of the IPMC, which is designed based on the shape of and efficient driving devices that can control the swing
a natural dolphin fin. The robot can jump and swim freely of the fishtail to achieve different amplitudes, different
like a dolphin (Shen et al. 2013) (Fig. 3d). The bionic frequencies, and different central positions and realize
robotic fish developed by the Chinese University of Hong the functions of acceleration, deceleration, and steering
Kong includes a rigid head, a wired-driven active body, of the released robot fish. Based on elastic plate defor-
and a flexible tail. A pair of SMA spring plates with the mation theory, the design size and motion input of the
same stiffness pass through an active body comprising elastic plate are inversely solved according to the motion
multiple connecting rods, which are like the backbone of function of the actual fish, which makes the deformation
a real fish, and then distribute a pair of wires along the motion of the elastic plate highly fit the fishtail swing in
spring plate to drive the moving body. The robotic fish reality. The 3D modeling and fluid simulation of the fish
tail is a flexible tail made of silicone and carbon fiber body were performed, and the geometric size and motion
reinforced material that allows the robotic fish to swim mode of the prototype were optimized. The bionic
in multiple modes, such as cruising, turning, rising, and robotic fish has good sealing properties in water and can
descending (Xie et al. 2020) (Fig. 3e). The robotic fish adjust its posture to achieve the flipping and pitching
designed by the New York University consists of a rigid functions. In 2019, the School of Mechanical Engineer-
acrylonitrile butadiene styrene (ABS) plastic body shell ing, Baicheng Normal University (Wang et al. 2019a)
and a tail consisting of rigid ABS elements and flexible proposed the concept of ’fundamental wave’, includ-
polyester tail fins. The robot fish uses a waterproof servo ing deformation description and linear density descrip-
motor to control the tail, and a flexible tail fin allows the tion, established the fish body wave model of the bionic
tail to bend and undulate to mimic the swimming of a live robotic fish, formed the control method system of the
fish. The tail beat frequency and amplitude of the robot multijoint bionic robotic fish’s stable swimming propul-
are controlled by an external microcontroller. The signals sion, and achieved the efficient and stable swimming of
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 7 of 28

the bionic robotic fish. In 2022, an underwater soft robot for the gait generation of robotic fish and developed a
was successfully developed by a joint team from the Max robot that uses a CPG for fish-like motion underwater.
Planck Institute for Intelligent Systems in Germany, These studies reveal that CPG-based approaches are
Seoul National University in South Korea, and Harvard easy to design, fast to implement, and capable of online
University in the United States (Ning et al. 2022). The adjustments (Chen et al. 2020) (Fig. 4b).
robot can swim underwater like a fish and automati- The CPG model includes four input parameters,
cally adjust its swing in the water according to the speed namely, flutter amplitude, flutter angular velocity, flutter
of the water. To design the controller for the robotic offset, and the time ratio of the beat phase to the recovery
fish, the research team developed a data-driven, lumped phase in the flutter. The robot fish developed by the South
parameter modeling method, which allows for accurate China University of Technology is equipped with three
but lightweight simulations using experimental data and infrared sensors installed on the left, front, and right
genetic algorithms, and the model can accurately predict sides of the robot fish, as well as an inertial measurement
the robotic fish’s behavior at drive frequency and pres- unit that can sense the surrounding obstacles and the
sure amplitude, including the effects of antagonistic co- direction of movement. Based on these sensor signals,
contraction on soft actuators (Li et al. 2023). Currently, CPG-based closed-loop control can drive the robotic
most of the simplified mathematical models are used for fish to avoid obstacles and track the specified direction
control. Still, the accuracy of the simplified mathemati- (Chen et al. 2021b) (Fig. 4c). The Peking University-
cal models is poor, and the robustness of the control sys- developed robotic fish uses CPG modeling as a nonlin-
tem is poor, which makes the underwater bionic robots ear oscillator for joints to realize coordination by altering
designed by this method have poor adaptability to the the connection weights between joints. The online gait
underwater environment. generation method based on CPG makes the transition
between swimming gaits elegant and smooth to real-
2.2 Central pattern generator (CPG)
ize multimode swimming and achieve a more realistic
The main control principle of the CPG is to utilize the movement. By changing the CPG parameters, various
mathematical model of the neuron network to drive the swimming patterns can be obtained to simulate the vari-
joint movement by imitating the movement law and ous movements of real fish in nature or designed based
biological control mechanism of the animal itself. The on special tasks (Zhao et al. 2006) (Fig. 4d). The Chinese
School of Intelligent Systems Science and Engineering Academy of Sciences (Yu et al. 2016) proposed a particle
at Harbin Engineering University (Wang et al. 2019b) swarm optimization (PSO)-based CPG control system
used four oscillators to construct a CPG network model for underwater vehicles. In general, the parameters of
to control the pectoral fin and tail fin with two degrees the CPG are determined manually based on experience
of freedom of multimode bionic robotic fish, which and computer numerical simulation. In this method, the
introduces the angle between the head and tail axis traveling wave parameters of robotic fish are given manu-
and the horizontal plane and the yaw angle as feedback ally, and 19 parameters, such as the optimal CPG connec-
information to control the swimming posture of the tion weight, self-inhibition coefficient, and time constant,
robotic fish and conducted in-depth discussion on the are selected through the PSO algorithm according to
motion control of the pectoral fin. The basic swimming the fish body wave equation. Simulation and experiment
strategy is developed based on Walker’s oscillating show the effectiveness of this method. The Hirose Labo-
pectoral fin model. Based on the multijoint robot fish ratory of Tokyo Institute of Technology (Nagai and Shin-
model, the National University of Singapore extracted take 2022) adopted the CPG control network comprising
two basic imitation swimming modes, ’cruise’ and this oscillator to control the robot, that is, a multijoint
’C-type sharp turn’, from the swimming observation of snake robot. The robot has 10 actuating units, constitut-
real fish as training samples. The general internal model ing a bilateral wave propulsion mechanism with bionic
imitates the CPG of the nervous system used to learn left and right counter muscles. The CPG control net-
and regenerate the coordinated behavior of fish. This work can generate rhythm joint angle control signals and
learning method can use general function approxima- achieve the yaw maneuvering of the robot. The simula-
tion capability and time/space scalability to generate tion test confirms the feasibility and effectiveness of the
the same or similar fish swimming patterns by adjust- control system (Alexander 2017). This control method
ing two parameters. The learned swimming mode was simulates the central nervous system well, generates con-
realized in the experiment of multiarticular robotic tinuous and coordinated control signals, and then gives
fish (Ren et al. 2013) (Fig. 4a). Waseda University built timely feedback to different environments. This method
and studied a CPG network with nonlinear oscillators is conducive to coordinated control and has a better envi-
ronmental adaptation effect, so it is widely used.
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 8 of 28

Fig. 4 CPG controls the bionic robot. a Robotic fish covered with waterproof tape swimming in the water (Ren et al. 2013); b Fabrication process
of the silicone tail and the outer view of the robotic fish (Chen et al. 2020); c Closed-loop CPG-based control can drive the robot fish (Chen et al.
2021b); d Prototype of multimode robotic fish (Zhao et al. 2006)

2.3 Sinusoidal controller control method rectangular tail fin. An air motor drives the first joint, and
A sine controller is a kind of control method that is the second is driven by a spring. A measurement system
widely used by researchers based on the fact that the is developed to measure the torque and angle of the first
waveforms and motion periods generated by how fish joint. The Polytechnic University of Milan (Bottasso et al.
are propelled are similar to sine functions. Thus, the sine 2008) successfully controlled a pair of pectoral fin joints
controller simplifies the motion process of fish into the and caudal-fin joints of a robotic fish by using a sinusoi-
frequency, amplitude, and waveform of the sine function dal controller and vibrator (a topology with three oscil-
and then controls the motion of each joint of the under- lators adjacent to each other). By movement of pectoral
water bionic robot through these parameters. At the and caudal joints, the robot can achieve various under-
same time, it changes the motion state by relying on the water swimming actions. The experimental results show
phase difference of the motion between different joints of that the control method can realize stable swimming.
the robot. The advantages of this control mode are sim- Due to the uniqueness of the function types in the con-
plicity and easy controllability. In 2015, the robot fish troller, this method has limitations. If there is a motion
’Pike’ was born at the Massachusetts Institute of Technol- mode that does not belong to the function characteris-
ogy; the hardware system of the robot fish ’Pike’ includes tics, it cannot be accurately regulated. In addition, this
a head, a pectoral fin, a tail fin, a dorsal fin, a main servo control mode has poor adaptability when dealing with
sine controller system, a pectoral fin servo system, and a the sudden change of control parameters, and it cannot
battery (Li 2015). In 2014, the Tokyo Institute of Tech- quickly adjust from one motion mode to another, leading
nology developed a self-propelled robot dolphin with two to poor environmental adaptability of the robot.
joints and an autonomous drive controller (Nakashima
and Karako 2014). The robot dolphin is a simplified 2.4 Rigid motor drive
model of a high-speed swimming marine creature with Most marine bionic robots are driven by rigid motors.
a length of 1.7 m, which is very close to the size of the Since motor-driven robots are easier to implement in
actual dolphin. The robot dolphin has a linear body and a terms of systems than flexible-driven robots, which can
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 9 of 28

fully use the high energy density and high efficiency of underwater robot is significantly enhanced. This series
motors, rigid motor-driven bionic robots are more con- of robots were applied to underwater archaeological dis-
venient for specific purposes (Karthik 2014). They are covery, experimental teaching, ocean cruise experiments,
currently more mature in development than flexible- and water quality detection and achieved good results.
driven bionic robots. For rigid motors, waterproof hous- The Harbin Engineering University-developed bionic
ing is often needed, with high sealing requirements and underwater robot is driven by two servo rigid motors
greater challenges in terms of water pressure (Dawson with tail fins and an interactive gear system, which can
and Allison 2020). achieve various complex movements, as well as two artic-
For bionic robots driven by rigid motors, there are ulated serpentines, HRF-I and HRF-II bionic robotic fish
mainly single-motor drives and multimotor drives. Mul- (Tian et al. 2022a). Compared with the former, the latter
timotor drive means that the system has more flexibility, can achieve steering, snorkeling, and reversing, and the
but there are more limitations regarding structure and performance in all aspects has been greatly improved.
size, and it can carry many functional sensors. Examples The Department of Computer Science at the University
include the UK Natural Environment Council’s (NERC) of Essex conducted experiments with a rigidly driven
5.5-m long 1800 kg-dry weight Autosub6000 AUV, which robotic fish G9 equipped with a variety of sensors and
is rated to a depth of 6000 m, can be equipped with found that it can respond to dynamic changes in its
a variety of payloads for marine geoscience research, environment, capturing its position in the tank and the
includes high-resolution multibeam echo sounders, sea- robot’s posture and internal state, with good drive perfor-
bed profilers, and side-scan sonar, color camera systems, mance (Liu and Hu 2006) (Fig. 5e). The New York Uni-
conductivity, temperature, depth, and electrochemical versity Institute of Technology designed a robot’s body
redox (Eh) sensors. It has precise navigation and terrain shell comprising a payload and a motor bay. The payload
tracking capabilities and has a sophisticated collision bay contains control electronics, batteries, and counter-
avoidance system (Wynn et al. 2014) (Fig. 5a). weights to enhance pitch and roll stability and achieve
The Robotics Institute of Beihang University success- appropriate buoyancy. More specifically, buoyancy is
fully developed the bionic robotic eel, bionic robotic set so the robot is almost completely submerged during
dolphin, experimental small robotic fish, and trail-tail operation. The cap provides a waterproof seal for the pay-
bionic robotic fish SPC-I, SPC-II, and SPC-III (Wang load bay and extends toward the rear of the robot, par-
et al. 2005) (Fig. 5b) driven by an electric motor and wire- tially covering the engine room. A toggle switch hidden
less remote-control rigid actuator (Li and Jiang 2012), as in the lid extension turns the robot on or off. The motor
shown in Figs. 5c and d. Compared with conventional compartment houses a Traxxas 2065 waterproof servo
motor-driven robots, the maneuverability of the bionic motor for the drive, which is connected to the rear by an

Fig. 5 Rigid motor-driven robot. a AUV submarine (Wynn et al. 2014); b SPC-I (Wang et al. 2005); c SPC-II (Liang et al. 2011); d Working environment
of SPC-III in the Taihu Lake (Liang et al. 2011); e G9 robotic fish profile (Liu and Hu 2006); f Top view of the robotic fish representing the robot’s
undulating tail (Kopman and Porfiri 2013)
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 10 of 28

improved servo motor horn. The caudal fin is snug in a buoyancy module, and a motor in the tail provides power
slit at the free end of the caudal fin (Kopman and Porfiri and adjusts its direction. The flexible part of the robotic
2013) (Fig. 5f ). fish, designed by the State Key Laboratory of Complex
Single motor-driven marine bionic robots are often Systems Management and Control at the Institute of
used in fish bionic robots, which have a single func- Automation, Chinese Academy of Sciences, consists of
tion and are not flexible enough in movement, such as three joints connected by an aluminum exoskeleton.
the bionic fish studied by Northern Research Center for Each joint is connected to an R/C servo motor that con-
Science and Technology at Malek Ashtar University of trols the rotation angle of the joint. The rubber caudal fin
Technology (Sabet and Nourmohammadi 2022) and the is connected to the third segment by the peduncle and is
voice-activated soft robot fish studied by Robert. The crescent-shaped with good coordination (Yu et al. 2016)
Massachusetts Institute of Technology (MIT) Distributed (Fig. 6c). The robot fish designed by the University of the
Robotics Laboratory developed a single-motor driven Chinese Academy of Sciences employs a magnetic actua-
robotic fish, a soft robotic fish system whose subcompo- tor as a motor. The propulsion system is characterized
nents include an elastomeric tail, an external gear pump, by remote control using Bluetooth low power and easy
two diving surfaces, and control electronics, including an operation through smart devices. By the electromagnetic
acoustic receiver and a fish eye-eye camera that can com- induction law, the robot fish can swim quickly and turn
plete underwater reconnaissance missions (Katzschmann flexibly. This miniature robot fish could be employed for
et al. 2018) (Fig. 6a). The robotic fish designed by the Col- animal behavior research and special underwater tasks
lege of Worcester employs a flexible body with embedded (Chen et al. 2017) (Fig. 6d).
rigid actuators that mimic the elongated anatomical form The Key Laboratory of Marine Engineering in Shan-
of a fish. Also, the robot has a novel fluid drive system dong Province developed a motor-driven robotic fish
that drives body movement and has all the subsystems of with artificial side-line sensors that can help enhance
traditional robots: power, drive, handling, and control. the fish’s maneuverability in dark environments. Arti-
A set of fluid elastomer actuators is at the heart of the ficial sidelines simulate the structure of fish sidelines,
fish’s soft body. The soft robot has an input-output rela- offering possibilities for underwater sensing technology
tionship similar to a biological fish, allowing it to be self- and robotic fish control (Salumäe and Kruusmaa 2013)
sufficient and capable of fast movement (Marchese et al. (Fig. 6e). Researchers from Shahrud University of Tech-
2014) (Fig. 6b). The Electrical Engineering and Com- nology studied a robotic fish with a streamlined drive
puter Science Department of the University of Michigan body and a flexible tail, comprising a network of pressur-
(Ozog et al. 2017). The robot adjusts its height through a ized liquid-filled chambers embedded in an elastic beam.

Fig. 6 Single motor-driven bionic robot. a Soft robotic fish and diver interface module (Katzschmann et al. 2018); b Details of a soft-bodied robotic
fish (Marchese et al. 2014); c Prototype of the robotic fish applied to the underwater robot competition (Yu et al. 2016); d Mechanical design
of the robotic fish (Chen et al. 2017); e FILOSE robot fish (Salumäe and Kruusmaa 2013); f Robot fish with a wire-driven active body and compliant
tail (Haji and Bamdad 2022)
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 11 of 28

Viscous fluids with different pressures flow in the chan- et al. 2012). Inspired by the amphibious tortoise, the
nel, producing normal and shear stresses in the channel, mother robot is designed with a spherical body, four
which can make the robot fish adapt to different water legs, and two degrees of freedom. Powered by 4 vec-
pressure environments (Haji and Bamdad 2022) (Fig. 6f ). tor water jets and 10 servo motors, it can walk on land
The salamander robot has a modular design compris- and cruise underwater (Shi et al. 2013) (Fig. 7d). The
ing seven drive elements and a head element (with the enhanced 3D printing, low cost, multifunction, high
same appearance as the others). The housing of each mobility, tortoise-like environmental monitoring, and
element includes two symmetrical parts molded with data acquisition mobile amphibious spherical robot by
a lightweight polyurethane resin. These components Beijing Institute of Technology has good amphibious
are connected using compatible connectors fixed to performance (Guo et al. 2018) (Fig. 7e). The cuttlefish
the output shaft. All output axes are aligned, so plane robot studied by the University of Nevada researchers
motion is produced. To ensure that the robot is water- is powered by two soft fins of multiple embedded IPMC
proof, a custom O-ring robot is used with a total length drive motors connected to an Eco-Flex membrane. The
of 77 cm. Asymmetric friction with the ground, which traveling wave is generated on the soft fin by drive, the
is required to crawl on the ground correctly, is achieved deformation and blocking force of IPMC on the soft fin
by fixing a pair of passive wheels to each element, thus are measured, and the actuator is characterized, which
ensuring a coordinated transition between swimming can have good wave swimming performance in water
and crawling of the robot (Crespi and Ijspeert 2008) (Shen et al. 2020) (Fig. 7f ).
(Fig. 7a). The four-legged starfish-shaped soft swim- Recently, bionic amphibious robots have developed
ming robot’s flexible and natural buoyancy offers many profoundly in bionic structure design, movement per-
advantages for tasks such as underwater exploration, formance, and outdoor workability. Researchers from
sample collection, and marine wildlife observations (Du Harbin Engineering University (Li et al. 2021) developed
et al. 2021) (Fig. 7b). In Fig. 7c, the swinging body of a shape-shifted bionic turtle that can travel in water and
the bionic robotic fish is a multilink mechanism con- walk on land. To enhance the reliability of bionic robots
nected by hinges and equipped with multiple motors. in the future, these robots designed for engineering appli-
In swimming, the required body curve can be acquired cations are driven by electric motors and are constantly
by adjusting the relative position of each connecting improved. However, due to their performance limita-
rod, optimizing its control performance and swimming tions, large size, and high power consumption, the size
efficiency compared with a single motor (Korkmaz and range of motors have become significant limitations.

Fig. 7 Multimotor driven bionic robot. a Salamander (left) and fish (right) robots (Crespi and Ijspeert 2008); b Soft starfish (Du et al. 2021);
c Demonstration of the body curve fitting method (Korkmaz et al. 2012); d Prototype of the spherical mother robot (Shi et al. 2013); e Amphibious
spherical robot (Guo et al. 2018); f Multimotor driven fish robot (Shen et al. 2020)
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 12 of 28

2.5 Soft actuator drive change from a bone-like motion to a Caran-shaped

For bionic robots driven by responsive soft actuators, motion (Praczyk 2014).
often by imitating the movement patterns of marine The Harvard University-designed completely soft octo-
organisms, artificial muscles are used to cause pro- pus robot has all parts of its body made by 3D print-
pulsion by deformation under control of voltage, and ing technology and feels slightly slimy to the touch. The
although their power and precision cannot be com- soft robot has morphing and cushioning and can travel
pared with those of electric motors, responsive soft through small, irregular spaces, which can be useful in
actuators are stimulated to perform better in terms of the medical, military, and exploration fields (Wehner
high adaptability due to their excellent compliance. et al. 2016) (Fig. 8a). The octopus robot, developed by
Meanwhile, soft actuator-driven bionic robots are Queen Mary University of London, is made entirely of
widely used in some miniature marine robots because soft materials and employs a new fluid drive mechanism
of their smaller size requirements due to the absence that allows the robot to push forward, change direction,
of motors, and soft actuators also have a huge advan- and rotate around its main axis. In addition, it can use
tage regarding range because they usually consume multiple tentacles to grab objects or propel them under-
less power when used compared to motor drives (Gao water (Fras et al. 2018) (Fig. 8b).
et al. 2022). In 2018, the Precision Engineering Institute The soft-bodied fish developed by Zhejiang University
designed a new robotic fish with an active and com- is powered entirely by a soft electroactive structure made
pliant propulsion mechanism, a maximum swimming of a DE and an ionic conducting hydrogel. The robot
speed of 2.15 body lengths per second, and a maxi- fish can swim at a speed of 6.4 cm/s (0.69 body length
mum instantaneous turning speed of 269°/s (Shintake per second), which is much faster than a soft responsive
et al. 2018). In 2014, Marche University of Technology material-powered previously reported soft robotic fish
designed a Carregi-shaped swimming robot through (Li et al. 2017) (Fig. 8c). In Fig. 8d, the flexible robotic
a multiphysics simulation environment, which can fish is a transitional stage between rigid robotic fish and

Fig. 8 Soft actuator-driven bionic robot. a Octopus robot (Wehner et al. 2016); b Multitentacle fish robot (Fras et al. 2018); c Soft electronic fish (Li
et al. 2017); d Whole body stiffness research (Chen and Jiang 2019)
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 13 of 28

flexible robotic fish, with typical soft materials and tra- Thus, in this review, a simplified classification model
ditional driving methods (Chen and Jiang 2019). Because for the robotics domain is put forward. First, similar to
the soft material has large elastic deformation, it can be BCF- and MPF-type motions in the fish classification,
restored to its original shape, and the soft material of robots can be classified as body or tail-end anal (BCA)
the robot fish can be used to protect the actuator and and mid-end or paired anal (MPA) (Wang et al. 2022).
waterproof (Liu and Jiang 2022). With the development BCA and MPA are classified based on where the drive
of bionics and materials science, marine release robots occurs relative to the central axis and the direction of
are increasingly driven by a variety of methods, and typi- robot propulsion. BCA achieves propulsion through
cal stimulus-responsive actuators include IPMC, SMA, drive motion along the central axis. However, in MPA,
responsive hydrogels, pneumatic structures, chemically the driving motion occurs outside the central axis. Simi-
responsive expanded fluid networks, and living cells (Bai lar to the fluctuating oscillation classification of fish, the
et al. 2021). Using simple principles and widely available subcategory is set to fluctuating oscillation motion based
materials, the highly integrated electric drive module not on the motion of the robot actuator. Like in the fish clas-
only eliminates bulky pumps, pipes, and other equipment sification, the fluctuating motion can be expressed as
but also enables precise control of deformation, while the the fluctuating motion in the actuator. In the same way,
compact form factor also increases portability. oscillatory motion refers to the propulsion structure that
rotates on its fixture instead of the wave-like shape.

3 Bionic robot drive mode and control strategy 3.1 Robot drive mode classification
Traditionally, underwater robots have been classified 3.1.1 MPF model
based on Breder’s fish classification: if a fish generates The movement modes of fish are classified by body parts
thrust by bending its body/or caudal fins, the resulting used by fish for propulsion into BCF propulsion mode
motion is classified as a BCF motion. Fish such as eel- and MPF propulsion mode (Zhou et al. 2023). In MPF
shaped, flesh-capsule, tuna, eel and shark can be catego- propulsion mode, the dorsal, ventral, pectoral, and anal
rized into BCF types (Jiao et al. 2022). However, suppose fins are mainly utilized as the main propulsion parts,
that fish uses their mid-fin (including anal, dorsal, and which can maintain high mobility, stability, and swim-
pectoral fins) or paired fins (including ventral and pecto- ming efficiency at low speed. In general, it can achieve
ral fins) as propulsion mechanisms, the resulting swim- accurate six degrees of freedom movement, underwa-
ming pattern is classified as a mid-fin or paired fin (MPF) ter position holding, and steering, but it is difficult to
movement (Wang and Wang 2014). Regardless of the achieve high-speed swimming and acceleration perfor-
mode of propulsion used, fish movements are character- mance is insufficient. Lampreys, an eel MPF swimming
ized by deformed bodies, fluid forces, and their interac- robot developed by the Marine Science Center of North-
tions. Moreover, each mode of motion can be classified eastern University, uses 10 TiNi filaments of 250 μm to
by wave frequency as fluctuating and oscillating, as can be energized and heated as a motor (Wu et al. 2013). It
be seen in fish movements. has a simple structure, no noise, and good stealth perfor-
Oscillatory motion can be applied when the fish gen- mance. The fishtail propulsion of the robot designed in
erates propulsion from wave-like motion. Otherwise, if this mode is quieter than the traditional propeller, which
the fish uses a rotation-like motion to obtain thrust, this is especially important in future naval battles. It can
motion can be classified as oscillatory. These two classi- greatly improve stealth ability. Japan developed the first
fications cannot be separated because oscillatory motion MPF robotic fish that can swim freely underwater; this
can be derived from the fluctuating motion of shorter bionic robotic fish is 650 mm long, 500 mm wide, and
wavelengths and vice versa. Eels, which usually use their 0.64 kg, with floating, diving, turn signal, and other func-
whole body to produce wave-like motions, can be clas- tions, and the smooth shape of the robot fish makes the
sified as fluctuating motions, while box fish, which only efficiency of the fishtail propeller up to 80% (Scaradozzi
make their tail fins swing due to body inflexibility, can be et al. 2017). It uses the three-joint bionic tail fin as the
classified as oscillatory motions. However, in this fish- only power source, with low power consumption, which
based classification, problems emerge in the general clas- can extend the battery life and is suitable for long-term
sification of animal species (Zhang et al. 2017). Especially underwater cruises, tracking, and other tasks. In 2021,
in robotics, there is no restriction to imitate the motion Osaka University developed a pair of miniature fish out
mechanism and shape of fish. Following the traditional of silicone, which was only the size of a hand but could
fish classification is still difficult if the motion mechanism swim at 0.1 m/s (Xie et al. 2021). In 2017, Festo, a Ger-
and shape of the robot are somewhat different from the man company, developed the pectoral fin bionic robotic
target animal. fish aqua ray with a body length of 615 mm, a wingspan
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 14 of 28

of 960 mm, and a maximum speed of 0.5 m/s (Saxena a BCF robotic fish (Wang et al. 2021). The width of the
and Chauhan 2017). The mechanical operation pro- first-generation bull nose shark robots is 28–46 cm, while
cess of forward, backward, differential turning, pitching, that of the second-generation one can reach more than
and other actions of the bionic robot was completed by 110 cm. The robot fish is driven by two motors on both
experiments and tests, completing the goal of the project. sides of the pectoral fin (60W DC motor drive); the flex-
In 2009, the China Academy of Automation developed ible pectoral fin comprises silicone rubber material, the
a small robot dolphin that is 560 mm long and weighs main material is made up of relatively low-density glass
about 3.3 kg, which can complete special tasks such as fiber, and the addition of gyroscopes can achieve auton-
marching, chasing, and searching (Xia et al. 2023). The omous navigation function. A steering engine drives
National University of Defense Technology in China pro- the BCF robot eel developed by the Beijing University
duced a prototype of a bionic pectoral fin powered by of Technology. The fin material is a composite material
multiple fins. In the water tank experiment, the robot’s that can be applied to the eddy current environment with
left and right fins moved simultaneously, with a forward a large water flow (Song et al. 2013). The robot shark,
speed of 0.13 m/s and a backward speed of 0.15 m/s. Due developed by St. Mary’s College, University of London,
to the symmetrical structure and movement of the wave also adopts the BCF drive mode (Watts and McGookin
fins of the robot, they could smoothly change the gait 2014). It is larger and adopts silicone fins, and the head’s
from forward to backward without turning and move lat- central processor controls the robot’s movement, which
erally by sending inward counterpropagation waves. can swim upstream in the rapids. The robot shark sim-
The Institute of Robotics of Beihang University and ulates a shark’s shape and swimming mode, with little
the Chinese Academy of Sciences successfully developed disturbance to the environment and no harm to under-
SPC-II bionic robotic fish, the first practical application water organisms. The multilink glider robot, developed
of MPF bionic robotic fish in China (Lou et al. 2017). by the China Academy of Electronics and Information
The SPC-II bionic robotic fish is 1.23 m long, with a Technology, can swim flexibly and glide efficiently in
shiny black body, a total weight of 40 kg, and a maximum 3D space and is equipped with the main BCF of a three-
diving depth of 5 m. It has a prominent camera above degree-of-freedom buoyancy drive system as the main
its head that collects location data. The SPC-II bionic propulsion device for stable propulsion in water (Wang
robotic fish can move, sink, and float freely and flexibly in et al. 2021) (Fig. 9a). A Lanzhou Jiaotong University-
waves. The MPF bionic robotic fish robot-ray I, robot-ray designed BCF-propelled four-fin bionic prototype based
II, robot-ray III, and robot-ray IV series were developed on modular design has high efficiency, rich turning
by Beihang University (Wu et al. 2021a). Among them, modes, good maneuverability, and high turning speed
the best-performing robot-ray IV is 320 mm long with a (Li et al. 2018) (Fig. 9b). Jilin University has developed
wingspan of 560 mm. The maximum swimming speed is a carpal bone robot fish with a four-degree-of-freedom
about 0.16 m/s. Moreover, the robot fish has high under- tail. The robot fish has two modes of radio control and
water speed and better load capacity, and the underwater autonomous swimming. The BCF mode has outstanding
movement trace is smaller. It can perform quick close-in performance of high speed and high efficiency (Yan et al.
intercepts, search for enemy divers, highly maneuverable 2008) (Fig. 9c). Ferat University in Turkey has developed
patrols, and track underwater targets at high speeds. The a bionic boat-shaped autonomous robotic fish prototype
bionic underwater vehicle based on the MPF long-wave with a double-link tail propulsion mechanism. To simu-
propulsion principle has the advantages of high mobil- late the robust swimming gait of fish, a bionic motion
ity, strong anti-interference ability, and environmental control structure based on CPG is adopted. The unidi-
friendliness (Korkmaz et al. 2015). Thus, research on this rectional chain CPG network designed is inspired by the
bionic underwater vehicle has a broad market prospect neural spinal cord of lampreys and is propelled by BCF.
and application value. It produces a steady, rhythmic pattern of oscillations
underwater (Ay et al. 2018) (Fig. 9d).
3.1.2 BCF model In 2015, the Harbin Institute of Technology success-
BCF propulsion mode enables most fish to swim by wav- fully developed a double-jointed Karan-shaped fish
ing or swinging part of the body and tail fin using eddy robot, code-named ’HRF-I’, with a swimming speed of
currents to push the water behind to use the water reac- 0.5 m/s (Wang et al. 2015). In 2018, the University of
tion force to achieve the forward movement of the fish Science and Technology of China designed a four-joint
body. When cruising at high speed, high swimming effi- bionic robotic fish based on the morphological structure
ciency can be achieved, generally, more than 80%, and and movement form of the Karan-shaped fish (Zhong
the acceleration and starting performance are good. The et al. 2018). In 2016, a BCF model bionic eel robotic fish
bionic bull nose shark designed by Beihang University is with eight joints was developed in the United States, and
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 15 of 28

Fig. 9 BCF mode swimming robot fish. a Main components of the FishBot (Wang et al. 2021); b Prototype of the proposed robotic fish (Li et al.
2018); c Mechanical structure of the robotic fish (Yan et al. 2008); d Detailed mechanical configuration of the robotic fish (Ay et al. 2018)

in 2017, Beihang University developed a series of fibular and thickness, such as turtles and penguins. It uses peri-
bionic robotic fish with two parallel joints in the tail stalk odic changes in the bending and sinking posture of the
and tail fin, driven by a two-axis servo motor (Yu et al. up and down swinging forelimbs to regulate the water’s
2017). The simulated or caudal-fin BCF pulsating under- angle of approach. It can be controlled independently
water thruster, developed by Osaka University in Japan in of the winging posture to produce forward thrust of the
2017, has flexible fins on both sides and is driven by 16 swim front itself, where the forelimbs swing in a process
DC servo motors via the top fins. The robot fish can real- that produces orthogonal directional (negative) drag and
ize flexible underwater movements such as surface, div- lift forces, prompting the body to keep advancing (Todd
ing, steering, pitching, and hovering, which confirms the et al. 2020). Although large, this biological body has the
viability of the application of undulation fin bionic under- advantages of explosive power, high efficiency, stabil-
water propellers to future underwater robots (Ravalli ity, low noise, excellent maneuverability, and operational
et al. 2017). Several flexible fish species, such as dolphins, performance. Several theoretical and experimental works
sharks, and tuna, swim in BCF mode and can swim with have been conducted on marine fluttering organisms by
high speed and efficiency. Based on this design, the BCF combining bionics from several disciplines. Based on
mode robot produces thrust by bending the torso and this, a series of bionic flutter wing propulsion devices
swinging the tail fin, leading to high swimming speed, have been developed with beneficial results. The Chinese
high efficiency, and fast starting performance; thus, the Academy of Sciences designed a four-joint robotic fish
BCF mode is suitable for applications such as long dis- head, which is a hollow, rigid, and streamlined shell made
tance and high-speed swimming, instantaneous accelera- of molded glass fiber that provides enough space for elec-
tion, or fast steering (Rajamohamed and Raviraj 2015). tromechanical components such as control circuits, sen-
sors, rechargeable batteries, and balancing heavy objects.
3.2 Bionic flutter rigid drive To duplicate the movement of fish, a series of multilink
The fluttering rigid drive mode is a structure of self- rigid motors connected by yaw joints are used as the main
excited vibration consisting of skeletal, muscular, and propulsion mechanism, followed by a slender tail shaft
nerve centers (Wang et al. 2021). It is the main mode of made of polyvinyl chloride and then a polyurethane tail
large aquatic animals with large spreading chord ratios fin with some elasticity. All the rods, driven by DC servo
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 16 of 28

motors, are connected in series to a metal skeleton cov- robot fish, NAF-I, weighs about 6.8 kg and is 650 mm
ered with a flexible waterproof skin that allows for flex- long, 100 mm wide, and 260 mm high. It is powered by
ible turns in water up to 213° (Su et al. 2014) (Fig. 10a). a 15 V nickel-metal hydride battery, allowing the fish
The robot adopts motor modularity to facilitate loading prototype to swim for up to 4 h when fully charged.
and unloading. In Fig. 10b, the Nanyang knife-fish robot One DC motor drives the oscillating tail fin, and the
contains three independent modules, namely, the buoy- other drives the counterweight, and the robot swims in
ancy box module, power cabin module, and wave fin a straight line at a speed of about 0.35 m/s, equivalent
module. Because these modules are designed in a mod- to about half a body length per second. It is also con-
ular manner, these modules can be easily replaced if the firmed that the greater the thrust of the motor on the
design changes or additional features need to be attached robot fish, the faster its swimming speed (Chong et al.
(Low 2009). 2009) (Fig. 10d). To produce greater thrust, the choice
To investigate the effects of electric motors on of motor parameters becomes very significant.
robotic fish, the University of Essex in the United Research on bionic flutter drive systems has never
Kingdom performed production tests on the G9 series stopped. Still, due to the complexity of the drive mech-
of robotic fish, which are about 52 cm long and have anism and its unique motion characteristics and the
three or four servo motors and two DC motors (Liu different research methods, the forms of flutter wing
and Hu 2010) (Fig. 10c). The servo motor is connected propulsion are also different (Zhu 2018). So far, it has
at the tail as three joints; the head is fitted with a DC been impossible to conduct a theoretical study for vari-
motor that changes the fish’s center of gravity, and ous bionic drive mechanisms because many crucial tech-
another controls a miniature pump that adjusts the nical and theoretical problems remain in the research
robot’s weight by pumping water. Enhancements in the stage, and the technical design of various bionic propul-
mechanical structure and skin materials have improved sion systems is still very backward and far from practical
the efficiency and robustness of the robotic fish. The application.

Fig. 10 Flutter rigid drive robot. a Slim fish robotic prototypes applied to C-start experiments (Su et al. 2014); b Southern Ocean knife-fish module
(Low 2009); c Schematic structure of a G9 series robotic fish (Liu and Hu 2010); d Prototype of biomimetic fish, NAF-I (Chong et al. 2009)
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 17 of 28

3.3 Bionic wave oscillation rigid drive robot with different modes of motion to adapt to the
At present, the main biomimetic fish propulsion systems complex marine environment. The motion mode of the
are BCF models, such as dolphins, which are propelled hybrid robot fish mainly has two types, namely, sail drive
by the caudal fin, and MPF models, such as manta rays, and wave drive. The HRF includes tail fins, wings, steer-
which are propelled by the pectoral fin. ing rubber, collapsible sails, and a hull. In wind-driven
The propulsion model has high thrust, stability, and mode, the sail is folded, while wave drive is used to drive
maneuverability (Jung et al. 2002). It has an excellent per- the hydrofoil up and down with waves to provide power;
formance in fast swimming under hydrostatic conditions thus, no extra energy is needed to move forward (Ma
and better start and stop functions but poor maneuver- et al. 2020) (Fig. 11b).
ability in low-speed turns and turbulent environments. In 2002, MIT developed the world’s first robotic fish—
Peking University developed a robot fish consisting of a bionic tuna—which can complete complex movements
rigid head, a flexible body, and a tail fin. The hard head such as propulsion, turning, and ascent diving (Koch
houses a control unit, a wireless communication mod- 2002). Its forward speed can reach 2 m/s, and the propul-
ule, and a set of batteries. The battery is placed at the sion efficiency is as high as 91%. Building on this, the MIT
bottom of the head to ensure the vertical stability of the team developed the reinforced fish Robopike and the
robot while swimming. A pair of pectoral fins are fixed steel-like underwater vehicle VCUUV in collaboration
on both sides of the head to ensure the stability of the fins with Draper Lab in the United States. In 2016, the birth
in water. The flexible body comprises three joints, each of these two robotic fish greatly improved the BCF mode
connected to a servo motor to adjust the deflection angle propulsion technology (Kumar et al. 2016). A hydrau-
of the joint. The rubber tail fin is fixed on the third joint lic bionic wave fin prototype is designed at the National
and acts with the water flow to move forward in waves University of Defense Technology, comprising a hydrau-
(Li et al. 2014a) (Fig. 11a). Developed by the Institute of lic pressure source, a hydraulic bionic wave fin principle
International Education, the HRF is a new type of marine prototype, and a data acquisition and processing system.

Fig. 11 Wave rigid drive robot. a Prototype of the robotic fish (Li et al. 2014a); b Prototype of the HRF in wind-driven mode (Ma et al. 2020); c BCF
mode swimming style (Chowdhury et al. 2011); d Amphibious snake robot (Kelasidi et al. 2016)
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 18 of 28

The flow variation rule, the function principle of bionic of them are only applicable to static laboratory environ-
oscillating joint movement, the underwater speed test, ments, and the working environment of bionic amphibi-
and the free navigation propulsion test were performed ous robots is complex and changeable. Thus, it needs to
on the prototype. The National University of Singapore sense the external environment information, parameter
developed a fish-like underwater vehicle integrating fish- change trend, and functional state in real time. In 2010,
like swimming, modular link, and fin movement. The the Tokyo Institute of Technology Robotics Laboratory
motor is used for simulation of the wave of the fish tail, designed the serpentine amphibious robot ACM-R5
that is, sinusoidal oscillation. The aim is to duplicate the based on the previously developed HELIX, which had
BCF model’s propulsion technology to swim efficiently poor performance (Yang and Ma 2010). The robot has
over long distances at impressive speeds (Chowdhury a 3D motion capability, and each module has a motion
et al. 2011) (Fig. 11c). Developed by the Norwegian Uni- mode of two degrees of freedom, capable of pitching and
versity of Science and Technology, the amphibious snake yawing. It has many gaits on land, but its gait in water
robot has similar kinematics whether on land or in water; has not been studied yet. To make serpentine amphibi-
the snake robot constantly changes its body shape to ous robots have more flexible mobility in water, the
reduce ground friction or hydrodynamic resistance to State Key Laboratory of Robotics of the Chinese Acad-
achieve forward propulsion, that is, when the snake robot emy of Sciences developed a new amphibious robot
follows a wavy gait pattern, it gains propulsion (Kelasidi called EXPLORER-III in 2020, which consists of nine
et al. 2016) (Fig. 11d). waterproof modular universal units, each with two free-
motion modes of pitch and yaw (Zheng et al. 2020). The
3.4 Special drive mode robot has a total length of 117 cm, a trunk diameter of
Various underwater organisms drive in different ways, 7.5 cm, and a total mass of 6.75 kg. Since 2016, the State
and simulation methods have always been employed Key Laboratory of Robotics of the Shenyang Institute
to explore their motion mechanism and optimize their of Automation, Chinese Academy of Sciences, has con-
motion to guide the design and production of underwa- ducted extensive research on another serpentine amphib-
ter robots. Several strange underwater organisms also ious robot and developed a prototype (Yang et al. 2016).
bring inspiration to researchers. For example, there is The robot is 700 mm long, 320 mm wide, and 150 mm
increasing research on underwater jellyfish octopus. The high, with a total mass of 4.995 kg. Moreover, the robot
organism is flexible, and only by changing the size of the can move at a speed of up to 0.45 m/s in water. Bionic
cavity does it achieve steering and fixed trajectory move- amphibious robots will simplify amphibious drive struc-
ment. However, its swimming stability is poor, and the tures by using soft actuators, improving energy efficiency,
direction is not easy to control, which is a huge issue to sensing the environment, and having a certain ability to
solve. Bionic water snake robots, like water snakes, can make autonomous decisions.
swim freely in water by swinging their tails. Their move- Wheel-propeller-integrated amphibious robots tend to
ment is flexible and can complete relatively complex task integrate multiple drive units, which can crawl in water
environments. They have good flexibility and freedom and on the ground. Thus, the driving device does not
in some locations that divers or other underwater vehi- need to be changed; only the mode of motion needs to be
cles cannot reach because of their appearance. Turtles changed, which can result in good motion performance
in water do not have the same slow movement as on the on land and in water (Liu and Jiang 2022). Thus, research
ground: they swim very fast and are very sensitive. Their on such robots worldwide has gradually increased,
unique way of propulsion also offers a lot of inspiration and researchers have achieved some great results. The
to researchers, especially those who study amphibious Mechanical Engineering and Automation major of Bei-
robot turtles. hang University designed an integrated wheel-propel-
ler amphibious robot, which has a simple and compact
3.4.1 Amphibian drive mode structure and can realize autonomous movement in two
Since the world’s first bionic amphibious robot was environments (Wu et al. 2021b). Shenyang Institute of
designed in 2013, it has gradually developed astonish- Automation, Chinese Academy of Sciences developed
ing achievements. The working environment of the an integrated wheel-propeller amphibious robot with
bionic amphibious robot consists of a beach, wetland, dimensions (L × W × H) of 1.0 m × 0.96 m × 0.2 m. The
underwater, and other complex terrains, and the bio- total weight is 44 kg, the maximum crawling speed is
logical prototype mainly comprises aquatic and ter- 1 m/s, the maximum swimming speed is 0.7 cm/s, and
restrial organisms. Researchers have established many the maximum working depth is 10 m (You et al. 2010).
theoretical models such as ’resistance theory’, ’slender Individual motors drive all drives of the robot, and
body theory’, and ’inverted pendulum model’, but most depending on the operating environment, the movement
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 19 of 28

can be easily switched by rotating the wheel-propel- of the liquid metal robot jellyfish integrating the interac-
ler 90°, but it needs to consume a lot of energy, and the tion of a fully flexible electromagnetic coil and a magnet
energy of bionic amphibious robots is extremely limited, for the first time, and designed a bionic robot jellyfish
and the efficiency of energy utilization is low, limiting its with more natural motion and propulsion (Zhou et al.
application. From the perspective of broadening income 2018) called RoMan-III. This is driven by a completely
sources and reducing expenses, on the one hand, bionic soft electromagnetic actuator, which can realize a vari-
amphibious robots must carry batteries with higher ety of soft swimming in response to different electrical
energy density and enhance outdoor energy collection signals. Based on further conceptual experiments and
capabilities. computational fluid dynamics simulations, Waseda Uni-
Revealing the movement characteristics of biologi- versity in Japan systematically explained the response
cal prototypes is the premise of bionic design. Due to mechanism of the robot jellyfish and various factors con-
the rapid development of biology, chemistry, structural trolling its movement behavior, including the formation
science, and other disciplines, research on the driv- of vortices and the way of rising, diving, and levitation,
ing mechanism of various underwater and land ani- and developed a bionic jellyfish with a spherical structure
mals has gradually entered the muscle tissue structure that can float better (Francis et al. 2002). By experiment,
and microcell energy utilization process. More accu- it was found that this structure can complete the retract-
rate mathematical models are required to offer a theo- able movement of jellyfish more smoothly.
retical basis for designing underwater and land-driven The squid water jet propulsion process principle is as
robotic structures. In bionic engineering science, several follows: First, the squid outer box membrane expands to
motion characteristics and swimming mechanisms of form negative pressure, and water fills the chamber. Sec-
aquatic organisms have not been fully explored, such as ond, the mantle shrinks sharply after the water jet and
the effect of dynamic instability on swimming efficiency funnel are closed. Finally, the air is rapidly ejected from
and the drag reduction function of aquatic organisms (Li the nozzle, and the body is subjected to a force in the
et al. 2021). Thus, there remains a big gap between most opposite direction of the airflow. The compressed shell
underwater bionic robots and their prototypes. Enhanc- of the stingray robotic fish, developed at Nanjing Univer-
ing the driving efficiency of wave motion and oscillation sity, is made of photosensitive resin, and the pectoral fin
motion is one of the crucial problems in solving bionic skeleton is composed of 12 carbon fiber rods. The robotic
wave motion, but so far, this problem has not been well fish uses a thin rubber film to squeeze the water around
solved. Thus, investigating the motion characteristics and it as it swims to generate thrust. The oscillation of fin
swimming mechanism of the bionic prototype and apply- rays causes the fluctuation of pectoral fins, and by con-
ing it to the bionic system, exploring the hydrodynamic trolling the amplitude, frequency, and phase difference
factors in the swimming process, and improving the between adjacent fins, different harmonic waveforms can
bionic similarity are the key issues to achieving efficient be produced (Wang et al. 2014b) (Fig. 12a). In Fig. 12b,
swimming of the model (Serhat, 2022). the bull nose fish robot simulates the pleural motion and
deformation of the bull nose rays. Each side of its internal
3.4.2 Bionic water jet soft drive skeleton comprises three fin-like rays, which are evenly
Aquatic cephalopods such as squid and jellyfish can con- distributed at the base of the fins along the chord. These
trol the contraction and expansion of the cavity through fins play a significant role in propulsion. The tail fin func-
muscle fibers during swimming, and their movement tions like a lifting rudder, producing power by beating the
is in an unstable state of acceleration and deceleration current to help the pectoral fin float and dive (Cai et al.
(Zhou et al. 2014). At the same time, they are propelled 2019).
by forces in the opposite direction of the water jet, which California Institute of Technology established a piston
enables mollusks such as jellyfish to move axially at jet model by studying the propulsion mechanism of a
extremely high instantaneous speeds and precisely posi- squid water jet (Wu et al. 2019) that used dynamic grid
tion themselves in slow motion. However, the expansion technology to simulate the formation process of vortex
and contraction of the cavity are not completed, and the rings under different spindle ratios and backgrounds.
air is slowly sucked in and out, leading to discontinuous The reasons for the formation of vortex rings were ana-
propulsion and poor movement continuity. From the lyzed, and the consistency of simulation and experimen-
above theories, the research group of the Liquid Metal tal results was effectively confirmed. Harbin Institute
Laboratory of the Institute of Physics and Chemistry of of Technology developed a water film and bionic noz-
the Chinese Academy of Sciences explored the motion zle based on a cuttlefish jet system (Tian et al. 2022b).
characteristics of jellyfish expansion and water absorp- The bending performance of the bionic nozzle was
tion, systematically discussed the theory and technology tested at different water temperatures and driving pulse
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 20 of 28

Fig. 12 Water spraying manta ray robot. a Robotic stingray design (Wang et al. 2014b); b Ox nose fish robot (Cai et al. 2019); c Robo-Ray IIs
(Kapetanovic et al. 2020); d Underwater robot with elastic skin (Ma et al. 2015)

conditions. Researchers used force sensors and high- on biomimetic water jet propulsion systems is still in its
definition cameras to capture and record the movement nascent stage: there is no relatively mature biomimetic
of the bionic jet system, effectively confirming the per- propulsion system, the types of technologies are rela-
formance characteristics of the bionic jet system (Wang tively small, there are several difficulties to be overcome,
et al. 2017). A new bionic manta ray robot was developed and there is a long way to go.
by Beihang University. The real flexible deformation of
pectoral fins can be well simulated by integrating flex- 4 Applications
ible mechanisms and rigid support into the mechanical Oceans are vital to life on Earth; they are key to regulat-
structure design of the robot. Second, the CPG con- ing climate and balancing various ecosystems (Park and
trol method is used to realize that the controller drives Kim 2016). They are also home to countless creatures
the rhythmic bionic movement, and the flapping wing and diverse environments. In addition, the oceans are
shoots water to push the body forward and up and down important channels for global transportation. They are
(Kapetanovic et al. 2020) (Fig. 12c). The bionic pectoral indispensable sources of energy. Despite their vital sig-
fin of the manta ray robot developed by Beihang Uni- nificance, oceans remain underexplored due to their
versity can produce an effective angle of attack, and the harsh conditions, making exploration impossible with
thrust generated by the interaction with the current can traditional methods. Using underwater vehicles for ocean
effectively propel the robot fish. The experimental results exploration is becoming increasingly popular as they
exhibit that the maximum forward speed of the robot fish allow people to conduct safe exploration in extreme envi-
can reach 0.43 m/s (0.94 times body length/second) when ronments for long periods. At present, underwater bionic
it is swimming in the tank, and it has good small radius robots are used in many fields, from oil and gas and fish-
turning maneuverability (Ma et al. 2015) (Fig. 12d). eries to archaeology, search, rescue, and defense (Li et al.
Due to different conditions, various bionic water jet 2014c). In addition, underwater robots are of use in sci-
propulsion systems cannot realize the same movement entific missions, such as mapping water composition and
as real organisms, nor do they have extremely sensitive environmental parameters over time and space, explor-
responses and fast movement ability. However, research ing the characteristics of the seafloor in terms of depth,
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 21 of 28

morphology, and composition, investigating glacial areas of underwater robots are continuing to improve, which
and icebergs, observing biological species in the envi- can better complete the work and is conducive to pro-
ronment, collecting biological and geological samples, moting large-scale development of the industry.
searching for life in the deep ocean, and helping protect
the environment from pollution. 4.2 Natural resource surveys
By duplicating the form of marine organisms, bionic
4.1 Application status of underwater robots robots can better adapt to harsh environments, such as
Since the second half of the twentieth century, under- high pressure, low temperature, and current, at the bot-
water robots have begun to assist human exploration tom of the sea. They are usually small in size and light in
of the ocean, and with the continuous advancement of weight; thus, they can better collect various substances in
human reach and exploration depth, underwater robots their original conditions, which is of great significance for
performing various tasks have also been born. In 2017, the study of natural resources at the bottom of the sea.
Professor Yang Canjun of Zhejiang University designed Underwater vehicles have been widely used in vari-
an underwater robot that can automatically clean marine ous marine geoscience research, initially focusing on
life 100 m below the surface of water. In its first sea test, seafloor mapping but more recently expanding to water
the robot sent back a ’selfie’ video underwater: firmly column and oceanographic surveys. The first underwater
attached to the wall of the tube, spraying water filled vehicle dedicated to marine was probably the IFREMER
with bubbles, and the accumulated shells were ’swept’ AUV, which was used in the early 1980s to map deep-
away. The robot is specially designed to clean the marine sea manganese nodule fields. A Woods Hole Oceano-
organisms attached to the surface of a steel pipe of an oil graphic Institution (WHOI) Sentry AUV was used to
drilling platform and has been successfully tested in the map the Deepwater Horizon oil spill in the Gulf of Mex-
Pinghu oil and gas field in the East China Sea. In 2018, ico, which resulted in a hydrocarbon plume (Levshonkov
the underwater unmanned robot enterprise Yoken Robot et al. 2020), using robots carrying detectors to assess its
launched a new product—BW Space Pro—which is the impact on animals and habitats. Many underwater vehi-
world’s first underwater UAV with intelligent functions, cles were deployed in 1995 and 1996 at the Juan de Fuca
which is widely used in diving entertainment, under- Ridge in the northwestern United States to detect and
water shooting, underwater survey, sea fishing, marine map new lava flows (Stenius et al. 2022). To use mag-
environmental protection, marine biological research, netometers to measure young lava flows at 2200 m east
aquaculture, underwater archaeology, underwater search longitude, WHOI developed the mixed-material under-
and rescue, and other fields. In 2019, Dr. Erik Engeberg water vehicle Nereus for scientific exploration at 11000 m
of Florida Atlantic University in the United States devel- in the deepest part of the ocean. This was almost twice
oped a jellyfish robot that can autonomically shuttle the depth range of the AUV at the time. In 2013, the
between coral reefs and monitor jellyfish robots at close French National Center for Marine Exploitation built
range. Besides assisting in research, the jellyfish robot Orca, an unmanned cable-free underwater vehicle with a
can shoulder the task of defending the ocean and serve maximum depth of 6000 m (Gao et al. 2013). In 2020, the
as a small spearhead in the front line of protecting the French National Sea Bomb Development Center cooper-
environment. In July 2020, the team of Professor Wen ated with a company to jointly develop the ’Eret’ acoustic
Li of Beijing University of Aeronautics and Astronau- remote-control diving robot, which is used for underwa-
tics and Junzhi, a researcher from the Institute of Auto- ter drilling rig inspection, submarine oil rig installation,
mation of the Chinese Academy of Sciences, designed oil pipeline auxiliary installation, anchor cable rein-
and manufactured an underwater soft robot arm that forcement, and other complex operations. In China, the
can be applied to the natural environment of the near underwater vehicle was first used in 2022 for subglacial
shallow sea, with the aim of establishing the kinematic surveys in the Arctic Ocean. Shortly after the scientific
model and rapid solution method of inverse kinematics survey ship ’Ocean’ began the third leg of the expedition,
to realize real-time kinematic control and finally realiz- ’Ocean’ conducted its first underwater robot operation
ing underwater grasp operations in the natural environ- in the East Pacific Sea for the first time with the under-
ment of the near shallow sea. With the upgrading and water robot ’Sea Dragon 2’, which was used to observe
mature application of underwater robot technology, it a rare giant chimney in the ’Bird’s nest’ black chimney
can not only greatly reduce the risk of manual operation area and carried a robotic arm used to accurately cap-
but also improve operation efficiency and reduce the cor- ture about 7 kg of vulcanized black chimney ventilation
responding expenditure cost. Meanwhile, driven by the samples. ’Hailong 2’ relying on accurate dynamic posi-
integration of other innovative technologies, both the tioning, accurately landed on the seabed in the black
comprehensive performance and cost performance levels chimney area of the ’Bird’s nest’ and performed camera
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 22 of 28

observation and measurement of hydrothermal environ- role in marine ecological protection. First, they can be
mental parameters. The discovery marks China as one used to collect marine environmental data. Using under-
of the few countries worldwide that can use underwater water robots, scientists can obtain detailed geographic
robots to conduct hydrothermal surveys and sampling images of the ocean and the conditions at the bottom
studies at mid-ocean ridges. The robot fish has the con- of the ocean. This data is crucial for understanding the
cealment of integrating into the fish, which can be used health and pollution levels of marine ecosystems. Pro-
to collect information on the fish or guide the fish to fessor Li Tiefeng’s team at Zhejiang University began
schedule the distribution or cluster of the fish according research on a bionic deep-sea soft robot based on lion-
to some algorithms (Marras and Porfiri 2012) (Fig. 13a). fish. Based on the dispersion and fusion of lionfish head
Thus, underwater bionic robots may effectively be used bones in soft tissues, the project team performed the
in marine environment observation, deep-sea resource mechanical design of the structure and material of elec-
exploration and development, and deep-sea and polar tronic devices and soft matrix and optimized the stress
scientific investigation. state in the robot body under a high-pressure environ-
ment. By designing materials and structures that adjust
4.3 Biodiversity research the devices and software, the robot could withstand a
Through the bionic robot’s similarity in appearance to deep-sea pressure of 10000 m without a pressure-resist-
marine life, marine life can be studied without disturb- ant shell and successfully conducted exploration missions
ing its normal activities, enabling close observation of in the Mariana Trench (Li et al. 2021) (Fig. 13b). Under-
marine life and potentially becoming a new platform for water robots can also be used to monitor the population
studying and interacting with underwater species (Wang and activity areas of marine life. With cameras and sen-
et al. 2020). Underwater bionic robots play an important sors, scientists can observe and record the behavior of

Fig. 13 Application of voice-activated soft machine fish. a Robot fish collect information about shoals of fish (Marras and Porfiri 2012); b Deep-sea
exploration (Li et al. 2021); c Underwater positioning (Wang et al. 2020); d Underwater imaging (Katzschmann et al. 2018)
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 23 of 28

many aquatic organisms in real time, providing evidence conventional underwater navigation equipment has limi-
for their conservation. tations; thus, bionic robots designed for different under-
In addition to data collection and monitoring, ROV water environments have great advantages.
maps can help protect marine life. They can remove The bionic underwater foot robot studied by the
debris and harmful substances from the ocean. Many National Metrology Institute of Japan (Maeda et al. 2020)
marine creatures often die by ingesting waste. The under- imitates the appearance and behavior of crabs and can
water vehicle can collect this waste through its robotic walk and jump underwater. Compared to traditional
arm and bring it to a safe location for disposal. Some AUV and ROV, it is better adapted to complex under-
underwater robot maps can even perform deep seabed water terrains and has a higher affinity for underwater
cleanup operations to help restore the health of the ocean organisms. Due to their bionic appearance, the natural
(Wang et al. 2002) (Fig. 13c). The University of Icahnx movements of underwater creatures can be well imaged.
developed a new kind of robotic fish for detecting pol- National Institute of Ocean Technology (Ramesh et al.
lution in river water and drawing 3D pollution maps 2017) used the bionic fish REMUS to map the habitat
of the river (Gomatam et al. 2012). Each robotic fish is at 1–2 m water depth in the Juan Strait in the northern
about 50 cm long, 15 cm high, and 12 cm wide. Each is United States. It used underwater video data for ground
equipped with pollution detection sensors and Global truth measurements. Underwater robots have been used
Positioning System (GPS), can ’smell’ harmful substances to map various seafloor morphological features, includ-
in the water, and can work together, even if there is no ing under ice sheets inaccessible to research ships. For
one to direct. When they ’sniff out’ the harmful sub- instance, State Marine Technical University (Siek and
stances in the water, they communicate with each other Sakovich 2019) used the underwater vehicle NERC Auto-
through a Wi-Fi wireless connection. The GPS navigation sub3 to investigate the retreat of the Pine Island Glacier
system allows them to swim freely without human opera- (PIG) in West Antarctica. The robot performed six mis-
tion, and once they find pollutants, they will send an alert sions in 94 h, collecting 510 km of orbital data under the
to the environmental protection department personnel PIG ice shelf 50 km above the ice surface.
(Skorohod et al. 2020). Underwater vehicles are also being used to image
Biosensors were first deployed on an underwater sedimentary features in submarine canyons. The Uni-
robot when an NERC autonomous submersible AUV versity of Kanagawa used an underwater vehicle car-
was fitted with an in situ dissolved manganese analyzer rying a high-resolution multibeam waveform acoustic
(Skorohod et al. 2020). This deployment showed how system (0.7 m lateral resolution) and a submarine pro-
an autonomous underwater robot carrying a biosensor filer (0.1 m vertical resolution) to conduct underwater
could detect small-scale changes in species distribution imaging experiments, collecting data from La Jolla Can-
that traditional sampling methods could not address. yon on the Southern California coast. To understand the
Since then, chemical sensors on underwater robots used processes that produce observational patterns on a scale
for marine purposes have been used mainly to search comparable to the surface (Tsukioka et al. 2002), the Sci-
the water column for active hydrothermal columns and ence and Technology on Underwater Vehicle Laboratory
to study species distributions, and a suite of sensors for used underwater robotic fish diving to collect deep-sea
detecting hazardous liquid spills have been deployed on data and provide vibration core samples for sediment
underwater robots in the North Sea Sleipner project for dating (Liu et al. 2020) (Fig. 13d). In the article on acous-
frequent, high time scale studies of areas of potential tically controlled soft robotic fish studied by the Uni-
spills to protect the ecological environment (Tran and versity of Zagreb (Kapetanovic et al. 2020), it is possible
Park 2020). By application of underwater robot mapping, to approach underwater organisms without disturbing
people can better protect the diversity of marine ecosys- their normal life and to image underwater organisms and
tems and marine life. They help people understand and underwater landscapes through shape features similar to
solve the problems of the marine environment and pro- those of fish (Katzschmann et al. 2018).
vide a guarantee for the rational use of marine resources. When conducting underwater imaging, conventional
underwater vehicles have higher accuracy in the tan-
4.4 Underwater imaging gential direction of the seabed and lower accuracy in
There is an increasing demand for exploration of the the vertical direction of the seabed. In comparison,
seabed environment, and the imaging requirements for underwater bionic robots have lower accuracy in the
marine resources and the underwater world are also tangential direction of the seabed and higher accu-
getting higher and higher (Liang et al. 2010). Due to the racy in the vertical direction of the seabed (Wang
uncertainty of the underwater environment, such as et al. 2014a). Moreover, the underwater bionic robot
interference of the current and limited sensing ability, has high stability and adaptability to the seabed
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 24 of 28

environment, and the combination of the two can 5 Summary and outlook
obtain higher-quality underwater imaging maps. From the above summarized research results, it can be
observed that research on bionic underwater robots has
grown considerably. Rapid turning, path tracking, auton-
4.5 Underwater search and rescue omous operation, and other actions have been achieved
Underwater robots can be used to check whether on some prototypes, and there is a great improvement in
explosives are installed on dams and bridge piers, speed and mobility, but there is still a very obvious gap
remote-reconnaissance structural conditions or dan- with real fish. Underwater bionic robot development is
gerous goods, and closely inspect underwater evi- high-end manufacturing industry supported by the Chi-
dence. In 2010, underwater robots could walk at nese government and plays the role of a ’strategic com-
3–6 km/h in the deepest underwater world of 6000 m manding height’. In China and abroad, a series of work
(Brown and Clark 2010). The forward-looking and has been conducted on the mechanical structure design,
downward-looking radar gives it ’good eyesight’. The materials, and control methods of underwater bionic
accompanying camera, video camera, and precise nav- robots, and the related research has grown considerably.
igation system allow it to ’overlook’. The underwater Due to the complexity of underwater, the mechanical
robot WHOI provided in 2012 took just a few days to structure design and control technology of underwater
find the wreckage of an Air France flight in 4000 ­km2 robots still require further optimization and improve-
of ocean after two years of fruitless searching by vari- ment to truly achieve a life-like system that integrates the
ous ships and aircraft. Underwater robots have great structure and biological characteristics. By enhancing the
potential and application value in rescue missions. characteristics of underwater robots, such as self-control
When encountering dangerous situations, underwater and self-perception, and through the coordinated control
bionic robots can play a greater role in on-site situa- of robot systems, underwater robots can better integrate
tion assessment and positioning, providing important into the underwater environment to complete the work,
information for the next step. Through the underwa- pursue sustainability on the road to development, and
ter high-definition camera group, sonar, and a variety make this technology more mature.
of sensors carried by the underwater robot itself, res- Research on bionic underwater robots has become
cue workers can grasp the water depth and tempera- more in-depth and expanded, and some prototypes have
ture on the shore. They can determine the obstacles in realized multimodal motion, fast turning, path tracking,
the water and remove the danger of entering the water autonomous operation, and other actions, which have
(Wang et al. 2019c). In salvage and other operations, greatly improved in speed and maneuverability. How-
the underwater robot can quickly locate the location ever, there is still a very obvious gap with real fish. In the
of underwater objects. Armed with this information, future, bionic underwater robots should be developed
commanders can better formulate a reasonable and into autonomous, intelligent, and collaborative tools. To
efficient rescue plan. Another major advantage of further improve the performance of the bionic under-
underwater robots lies in search and rescue missions. water robot system, further work should be conducted
In dangerous waters such as rapids and low tempera- in the following main research directions: (1) Mecha-
tures, it can take the lead in entering underwater areas nism design and optimization. Most bionic underwater
that rescuers cannot reach to detect the location and robots are driven by motors. Research can be conducted
situation of trapped people. The robot operator can in terms of streamlined low-resistance shapes, intelligent
control the movement of the robot by manipulating driving materials, and rigid and flexible coupling efficient
the handle or wireless sensing device on the shore. transmission mechanisms to improve the motion per-
Carrying tools such as robotic arms can also assist formance of bionic underwater robots. (2) Underwater
rescue workers in completing tasks such as clearing environment perception and modeling are significant for
and salvaging. The water environment where the dan- bionic underwater robots to perform underwater tasks.
ger occurs is not always ideal, and low visibility is one Information fusion technology of various sensors can
of the most significant problems. Bionic fluorescent be examined and combined with the technology to con-
robot fish can provide rescuers with a light source, and duct underwater environment modeling and improve the
rescuers can also determine the location of the target autonomy intelligence of bionic underwater robots. (3)
and search for risks by referring to the umbilical cable Intelligent control methods, such as artificial intelligence,
connected to the underwater robot (Asadnia et al. are a hot field right now. Some artificial intelligence
2015). The emergence of underwater robots makes technologies, such as reinforcement learning and trans-
underwater rescue work safer and more efficient. fer learning, can be applied to the intelligent control of
bionic underwater robots so that they can learn various
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 25 of 28

motor skills independently. (4) Multibionic underwater Declarations

robot cooperation. In nature, fish is often in the form of
Ethics approval and consent to participate
clusters for foraging, defense, and cruising. The use of Not applicable.
multiple bionic underwater vehicles to form a coopera-
tive system is helpful in improving operational efficiency. Consent for publication
Not applicable.
Due to the complexity of underwater, the particularity of
the propulsion mechanism, and the bottleneck of under- Competing interests
water communication, sensing, positioning, and other The author declares that there are no conflicts of interest regarding the publi-
cation of this paper.
technologies, the collaboration of multibionic underwa-
ter robots will be a very challenging direction.
Due to the complexity of the marine environment, Received: 24 August 2023 Revised: 12 September 2023 Accepted: 14
September 2023
underwater bionic robots will face problems such as
the drastic change in water velocity, the difference in
pressure under different water depths, and their water-
proofing, which poses a great challenge to the structural References
design of robots. To address these issues, the structure Alexander P (2017) Robot fish: bio-inspired fishlike underwater robots. Under-
of the future underwater bionic robot needs to be more water Technol Int J Soc Underwater 34(3):143–145
Asadnia M, Kottapalli AGP, Haghighi R, Cloitre A, Alvarado PV, Miao JM et al
detailed and more lightweight, and the application of (2015) MEMS sensors for assessing flow-related control of an underwa-
materials should also meet the requirements of the ter biomimetic robotic stingray. Bioinspir Biomim 10(3):10
underwater environment. Miniaturization is the cur- Ay M, Korkmaz D, Koca GO, Bal C, Akpolat ZH, Bingol MC (2018) Mechatronic
design and manufacturing of the intelligent robotic fish for bio-inspired
rent trend of robot development because small struc- swimming modes. Electronics 7(7):118
tures are easier to adapt to the environment, reduce the Bai XJ, Wang Y, Wang S, Wang R, Tan M, Wang W (2021) Modeling and analysis
contact area, and thus reduce the impact of underwater of an underwater biomimetic vehicle-manipulator system. Sci China Inf
Sci 65(3):35
pressure on the machine structure to a greater extent. Bal C, Koca GO, Korkmaz D, Akpolat ZH, Ay M (2019) CPG-based autonomous
The most prominent point is that miniaturized robots swimming control for multi-tasks of a biomimetic robotic fish. Ocean
are closer to the physiological structure of marine organ- Eng 189:106334
Bottasso CL, Scorcelletti F, Ruzzene M, Ahn SS (2008) Trajectory optimization
isms and fundamentally realize the bionic effect rather for DDE models of supercavitating underwater vehicles. J Dyn Syst
than just the imitation of appearance. Marine space is Meas Control 131(1):18
generally unsuitable for human survival, and large-scale Brown C, Clark RP (2010) Automated conceptual design utility for unmanned
underwater vehicles. Sea Technol 51(12):33–36
development and utilization of marine resources have Cai YR, Bi SS, Li GY, Hildre HP, Zhang HX (2019) From natural complexity to
a great dependence on robotics technology. Replacing biomimetic simplification: the realization of bionic fish inspired by the
humans with robots to promote and realize unmanned Cownose Ray. IEEE Robot Autom Mag 26(3):27–38
Cao Q, Wang R, Zhang T, Wang Y, Wang S (2022) Hydrodynamic modeling
marine equipment has far-reaching strategic significance. and parameter identification of a bionic underwater vehicle: RobDact.
Thus, future bionic underwater robots should be further Cyborg Bionic Syst 2022:13
developed, mainly in the direction of autonomy, intelli- Chen AL, Thind K, Demir KG, Gu GX (2021a) Modeling bioinspired fish scale
designs via a geometric and numerical approach. Materials 14(18):5378
gence, and synergy, to enhance the performance of bionic Chen BX, Jiang HZ (2019) Swimming performance of a tensegrity robotic fish.
underwater robotic systems. Soft Rob 6(4):520–531
Chen JY, Yin B, Wang CC, Xie FR, Du RX, Zhong Y (2021b) Bioinspired closed-
Acknowledgements loop CPG-based control of a robot fish for obstacle avoidance and
This work was supported by grants from the National Natural Science Founda- direction tracking. J Bionic Eng 18(1):171–183
tion of China (Grant Nos. 62201537 and U20A20194), the Natural Science Chen X, Shigemune H, Sawada H (2020) An untethered bionic robotic fish
Foundation of Shandong Province (Grant No. ZR2022QF008), and the Central using SMA actuators. Int Conf Mechatr Auto 13(6):1768–1773
University Basic Research Fund of China (Grant No. 202312035). We thank Chen X, Wu Z, Yu J, Zhou C, Yu JZ (2017) Design and implementation of a mag-
Yanyue Teng at the Ocean University of China for the useful discussion on the netically actuated miniature robotic fish. 20th World Congress of the
soft actuator drive part of this review. International-Federation-of-Automatic-Control (IFAC), Toulouse, France,
9-14 July 2017, pp 6851–6856
Authors’ contributions Chong CW, Zhong Y, Zhou CL, Low KH, Seet GLG, Lim HB (2009) Can the
Zhongao Cui and Liao Li performed the literature survey, drafted the manu- swimming thrust of BCF biomimetics fish be enhanced. 2008 IEEE Inter-
script and revised it critically for the key content. Yuhang Wang conducted national Conference on Robotics and Biomimetics, Bangkok, Thailand,
literature research and content verification. Zhiwei Zhong carried out the 22-25 February 2009, pp 437–442
document sorting and figure modification. Junyang Li is the correspond- Chowdhury AR, Prasad B, Kumar V, Kumar R, Panda SK (2011) Design, modeling
ing author, responsible for organizing the manuscript sequence alignment, and open-loop control of a BCF mode bio-mimetic robotic fish. Int
proofreading and revising the manuscript, and giving the final approval of the Siberian Conf Contr Commun 15(3):87–93
version to be published. All authors read and approved the final manuscript. Chu WS, Lee KT, Song SH, Han MW, Lee JY, Kim HS et al (2012) Review of
biomimetic underwater robots using smart actuators. Int J Precis Eng
Availability of data and materials Manuf 13:1281–1292
The data and materials used to support the findings of this study are included Crespi A, Ijspeert AJ (2008) Online optimization of swimming and crawling in
in the article. an amphibious snake robot. IEEE Trans Rob 24(1):75–87
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 26 of 28

Dawson HA, Allison M (2020) Requirements for Autonomous Underwater Vehi- Kumar S, Rastogi V, Gupta P (2016) Recent developments in modeling
cles (AUVs) for scientific data collection in the Laurentian Great Lakes: a and control of underwater robot manipulator: a review. Indian J Sci
questionnaire survey. J Gt Lakes Res 47(1):259–265 Technol 9(48):8
Du T, Hughes J, Wah S, Matusik W, Rus D (2021) Underwater soft robot mod- Levshonkov NV, Gusev AL, Krylosova AA (2020) Calculation of vibrations in
eling and control with differentiable simulation. IEEE Robot Auto Lett a tethered underwater vehicle-umbilical cable system. J Mach Manuf
6(3):4994–5001 Reliab 49(7):562–567
Francis O, Coakley G, Kitts C (2002) A digital control system for the triton Li DJ (2015) Analysis on numerical simulation for bionic robot fish based on
undersea robot. IFAC Proc Vol 35(2):633–638 CATIA. Appl Mech Mat 742:511–515
Fras J, Noh Y, Macias M, Wurdemann H, Althoefer K (2018) Bio-inspired octopus Li G, Chen X, Zhou F, Liang Y, Xiao Y, Cao X et al (2021) Self-powered soft
robot based on novel soft fluidic actuator. IEEE Int Conf Robot Auto robot in the Mariana trench. Nature 591(7848):66–71
21(5):1583–1588 Li J, Zhang YX, Li WB (2021) Formation control of a multi-autonomous
Gao CZ, Du YC, Yang L (2022) The influence of the discharge port structure on underwater vehicle event-triggered mechanism based on the Hun-
the infrared characteristics of underwater vehicle thermal jets. Appl Sci garian algorithm. Machines 9(12):346
12(14):7108 Li L, Wang C, Xie GM (2014a) Modeling of a carangiform-like robotic fish for
Gao JW, Yu JP, Li QC, Leng WZ, Ma ZT (2013) A layered obstacles avoid- both forward and backward swimming: based on the fixed point.
ance algorithm for biomimetic robotic fish. ICIC Express Letters IEEE International Conference on Robotics and Automation (ICRA),
7(1):1798–1810 Hong Kong, pp 800–805
Gomatam S, Vengadesan S, Bhattacharyya SK (2012) Numerical simulations of Li T, Li G, Liang Y, Cheng T, Dai J, Yang X et al (2017) Fast-moving soft elec-
flow past an autonomous underwater vehicle at various drift angles. J tronic fish. Sci Adv 4(1):7
Nav Archit Mar Eng 9(2012):135–152 Li WK, Chen H, Cui WC, Song CH, Chen LK (2023) Multi-objective evolution-
Guo S, He Y, Shi L, Pan S, Xiao R, Tang K et al (2018) Modeling and experimental ary design of central pattern generator network for biomimetic
evaluation of an improved amphibious robot with compact structure. robotic fish. Complex Intell Syst 9:1707–1727
Robot Comp-Integr Manufact 51(2):37–52 Li Y, Jiang YQ (2012) The application of distributed multi-sensor infor-
Guo SX, Shi LW, Ye XF, Li LF (2007) A new jellyfish type of underwater micro- mation fusion technology in underwater vehicle. Adv Mat Res
robot. 2007 International Conference on Mechatronics and Automa- 532–533:1006–1010
tion, Harbin, pp 509–514 Li Y, Jiang YQ, Ma S, Chen PY, Li YM (2014b) Inverse speed analysis and
Haji BJ, Bamdad M (2022) Nonlinear modeling and analysis of a novel robot low speed control of underwater vehicle. J Centr South Univ
fish with compliant fluidic actuator as a tail. J Bionic Eng 19(3):629–642 21(7):2652–2659
Hu Q, Dong E, Cheng G, Jin H, Yang J, Sun D (2019) Inchworm-inspired soft Li YS, Sun HX, Chu M, Zhang YH, Jia QX, Lan XJ (2014c) Experiment, simula-
climbing robot using microspine arrays. IEEE/RSJ Int Conf Intell Robots tion and analysis on coupling hydrodynamic forces under key param-
Syst (IROS) 78(5):5800–5805 eters for a spherical underwater exploration robot. J Vibroengine
Hu QQ, Dong EB, Sun D (2023) Soft modular climbing robots. IEEE Trans Robot 16(6):3014–3025
39(1):399–416 Li ZG, Ge LM, Xu WQ, Du YJ (2018) Turning characteristics of biomimetic
Hu QS, Zhou H (2009) IPMC propelled biomimetics robotic fish energy robotic fish driven by two degrees of freedom of pectoral fins and
consumption model construction and its application to energy-saving flexible body/caudal fin. Int J Adv Rob Syst 15(1):21
control. IEEE International Conference on Robotics and Biomimetics Li ZJ, Yin PL, Jiang XM, Tang LH, Wu H, Peng Y et al (2021) Towards self-
(ROBIO), Guilin, China, 19-23 December 2009, pp 2151–2156 powered technique in underwater robots via a high-efficiency elec-
Jiao ZW, Wang HY, Luo B, Yang WM, Yu Y (2022) A BCF bionic robot fish driven tromagnetic transducer with circularly abrupt magnetic flux density
by a dielectric elastomer actuator. J Phys: Conf Ser 2331(1):12010 change. Appl Energy 302(4):793–798
Jung KY, Kim, IS, Yang SY, Lee MH (2002) Autopilot design of an autonomous Liang JH, Wang TM, Wen L (2011) Development of a two-joint robotic fish
underwater vehicle using robust control. Int J Control Autom Syst for real-world exploration. J Field Robot 28(1):70–79
4(4):264–269 Liang XA, Zhang JD, Li W, Su LF (2010) Extended Kalman filter based
Kapetanovic N, Vasilijevic A, Nad D, Zubcic K, Miskovic N (2020) Marine robots identification of dynamic model for underwater robots. International
mapping the present and the past: unraveling the secrets of the deep. Conference on Applied Mechanics and Mechanical Engineering,
Remote Sens 12(23):3902 Changsha, China, 8-9 September 2010, pp 780–783
Karthik S (2014) Underwater vehicle for surveillance with navigation and Liu G, Wang M, Xu L, Incecik A, Sotelo MA, Li Z et al (2020) A new bionic
swarm network communication. Indian J Sci Technol 7(S6):22–31 lateral line system applied to pitch motion parameters perception for
Katzschmann RK, DelPreto J, MacCurdy R, Rus DL (2018) Exploration of under- autonomous underwater vehicles. Appl Ocean Res 99:102142
water life with an acoustically controlled soft robotic fish. Sci Robot Liu J, Hu H (2010) Biological inspiration: from carangiform fish to multi-joint
3(1):12–22 robotic fish. J Bionic Eng 7(1):35–48
Kelasidi E, Pettersen KY, Liljeback P, Gravdahl JT (2016) Locomotion efficiency Liu JD, Hu HS (2006) Biologically inspired behaviour design for autonomous
of underwater snake robots with thrusters. 14th IEEE International robotic fish. Int J Autom Comput 3(4):336–347
Symposium on Safety, Security, and Rescue Robotics (SSRR), Lausanne, Liu YW, Jiang HZ (2022) Research development on fish swimming. Chin J
Switzerland, 23-27 October 2016, pp 174–181 Mech Engine 35(1):114
Khalaji AK, Zahedifar R (2020) Lyapunov-based formation control of underwa- Lou BD, Ni YJ, Mao MH, Wang P, Cong Y (2017) Optimization of the kinematic
ter robots. Robotica 38(6):1105–1122 model for biomimetic robotic fish with rigid headshaking mitigation.
Ko Y, Na S, Lee Y, Cha K, Ko SY, Park J et al (2012) A jellyfish-like swimming mini- Robotics 6(4):30
robot actuated by an electromagnetic actuation system. Smart Mater Low KH (2009) Modelling and parametric study of modular undulating fin
Struct 21(5):057001 rays for fish robots. Mech Mach Theory 44(3):615–632
Koch RM (2002) Analysis of variable-thickness, streamlined transducer array Ma HW, Cai YR, Wang YL, Bi SS, Gong Z (2015) A biomimetic cownose ray
windowing concepts for high speed underwater vehicles. J Acoust Soc robot fish with oscillating and chordwise twisting flexible pectoral
Am 112(5):2217–2218 fins. Industr Robot: Int J 42(3):214–221
Kopman V, Porfiri M (2013) Design, modeling, and characterization of a Ma LP, Yue ZJ, Zhang RF (2020) Path tracking control of hybrid-driven
miniature robotic fish for research and education in biomimetics and robotic fish based on deep reinforcement learning. 2020 IEEE
bioinspiration. IEEE/ASME Trans Mechatron 18(2):471–483 International Conference on Mechatronics and Automation (ICMA),
Korkmaz D, Akpolat ZH, Soyguder S, Alli H (2015) Dynamic simulation model Beijing, China, 13-16 October 2020, pp 815–820
of a biomimetic robotic fish with multi-joint propulsion mechanism. Ma YJ, Li XY, Liang B, Du Y, Liu JQ, Yang S (2019) Development and motion
Trans Inst Meas Contr 37(5):684–695 mode analysis of IPMC bionic jellyfish based on app bluetooth
Korkmaz D, Budak U, Bal C, Koca GO, Akpolat Z (2012) Modeling and imple- remote control. IOP Conf Ser: Earth Environ Sci 252(2):022040
mentation of a biomimetic robotic fish. Int Symp Power Electr Power
Electr Electr Drives, Automat Motion 77(5):1187–1192
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 27 of 28

Maeda Y, Kida Y, Deguchi M (2020) Hadal free-fall camera lander transmit- Shi LW, Guo SX, Mao SL, Yue CF, Li MX, Asaka K (2013) Development of an
ting hadal images via high-rate acoustic communication. Sea Technol amphibious turtle-inspired spherical mother robot. J Bionic Eng
61(12):22–24 10(4):446–455
Marchese AD, Onal CD, Rus D (2014) Autonomous soft robotic fish capable Shintake J, Cacucciolo V, Shea H, Floreano D (2018) Soft biomimetic
of escape maneuvers using fluidic elastomer actuators. Soft Robot fish robot made of dielectric elastomer actuators. Soft Robot
1(1):75–87 5(4):466–474
Marras S, Porfiri M (2012) Fish and robots swimming together: attraction Siek Y, Sakovich S (2019) Simulation of the controlled movement based
towards the robot demands biomimetic locomotion. J R Soc Interf on the complexity principle for an automatic underwater vehicle.
9(73):1856–1868 Vibroeng Procedia 25:194–200
Nagai T, Shintake J (2022) Rolled dielectric elastomer antagonistic actuators for Skorohod BA, Statsenko AV, Fateev SI, Zhilyakov PV (2020) Accuracy analysis
biomimetic underwater robots. Polymers 14(21):4549 of 3D points reconstructed from workspace of underwater robot. J
Najem J, Sarles SA, Akle B, Leo DJ (2012) Biomimetic jellyfish-inspired under- Phys: Conf Ser 1661(1):1–7
water vehicle actuated by ionic polymer metal composite actuators. Song TL, Lu YP, Liu HQ (2013) Control and research of bionic robotic fish
Smart Mater Struct 21(9):11 based on Arduino. Adv Mat Res 2428(706–708):691–694
Nakashima M, Karako Y (2014) Effect of bubbles around an underwater Stenius I, Folkesson J, Bhat S, Sprague CI, Ling L, Özkahraman Ö et al (2022)
robot arm on thrust during the crawl stroke motion. Proc Engine A system for autonomous seaweed farm inspection with an under-
72(1):715–719 water robot. Sensors 22(13):5064
Ning KW, Hartono P, Sawada H (2022) Using inverse learning for controlling Su ZS, Yu JZ, Tan M, Zhang JW (2014) Implementing flexible and fast turning
bionic robotic fish with SMA actuators. MRS Adv 7(30):649–655 maneuvers of a multijoint robotic fish. IEEE/ASME Trans Mechatron
Ozog P, Johnson-Roberson M, Eustice RM (2017) Mapping underwater ship 19(1):329–338
hulls using a model-assisted bundle adjustment framework. Robot Suk YT, Hwan KM (2014) Analysis of integrated navigation performance for
Auton Syst 87(4):329–347 sensor selection of Unmanned Underwater Vehicle (UUV). J Ocean
Park J, Kim J (2016) Model-based pose estimation for high-precise underwater Engine Technol 28(6):566–573
navigation using monocular vision. J Korea Robot Soc 11(4):14 Tian QH, Wang T, Wang YX, Li CJ, Liu B (2022a) Robust optimization design
Pham TH, Nguyen K, Park HC (2023) A robotic fish capable of fast underwater for path planning of bionic robotic fish in the presence of ocean cur-
swimming and water leaping with high Froude number. Ocean Eng rents. J Marine Sci Engine 10(8):1109
268(1):113512 Tian QH, Wang T, Wang YX, Wang ZW, Liu CW (2022b) A two-level opti-
Praczyk T (2014) Using augmenting modular neural networks to evolve neuro- mization algorithm for path planning of bionic robotic fish in the
controllers for a team of underwater vehicles. Soft Comput Fusion three-dimensional environment with ocean currents and moving
Foundations Methodol Appl 18(12):2445–2460 obstacles. Ocean Eng 266(P3):112829
Rajamohamed S, Raviraj P (2015) Bio-inspired swimming techniques for Todd C, Samuel L, Yahya MS (2020) A bio-inspired robotic fish utilizes the
robotic fish using fow and pressure sensing mechanism (computa- snap-through buckling of its spine to generate accelerations of more
tional bio-mimetic). Ind J Sci Technol 8(24):10 than 20g. Bioinspir Biomim 15(5):1–18
Ramesh S, Ramadass GA, Prakash VD, Sandhya CS, Ramesh R, Sathianarayanan Trabia S, Shen Q, Stalbaum T, Hunt R, Hwang T, Kim K (2016) Numerical and
D et al (2017) Application of indigenously developed remotely oper- experimental investigation of a biomimetic robotic jellyfish actuated
ated vehicle for the study of driving parameters of coral reef habitat of by ionic polymer-metal composite. 13th International Conference
South Andaman Islands India. Curr Sci 113(12):2353–2359 on Ubiquitous Robots and Ambient Intelligence (URAI), Xi’an, China,
Ravalli A, Rossi C, Marrazza G (2017) Bio-inspired fish robot based on chemical 19-22 August 2016, pp 204–205
sensors. Sens Actuators, B Chem 239:325–329 Tran LG, Park WT (2020) Biomimetic flow sensor for detecting flow rate
Ren QY, Xu JX, Fan LP, Niu XL (2013) A GIM-based biomimetic learning and direction as an application for maneuvering autonomous
approach for motion generation of a multi-joint robotic fish. J Bionic underwater vehicle. Int J Prec Engine Manufacturing-Green Technol
Eng 10(4):423–433 9(1):163–173
Rossi C, Colorado J, Coral W, Barrientos A (2011) Bending continuous structures Tsukioka S, Aoki T, Ochi H, Shimura T, Sawa T, Nakamura T et al (2002) Devel-
with SMAs: a novel robotic fish design. Bioinspir Biomim 6(4):045005 opment of an acoustic lens for an imaging sonar for autonomous
Sabet M, Nourmohammadi H (2022) Water velocity sensor with the ability to underwater vehicle ’Urashima’ and experimentation in a water tank. Jpn
estimate the sideslip angle based on Bernoulli’s law for use in autono- J Appl Phys 41(6R):1–10
mous underwater vehicles. Ocean Eng 263(3):112252 Villanueva A, Smith C, Priya S (2011) A biomimetic robotic jellyfish (Robo-
Şafak KK, Adams GG (2002) Dynamic modeling and hydrodynamic perfor- jelly) actuated by shape memory alloy composite actuators. Bioinspir
mance of biomimetic underwater robot locomotion. Auton Robot Biomim 6(3):036004
13(3):233–240 Vogel V (2012) Soft robotics: bionic jellyfish. Nat Mater 11(10):841–842
Salumäe T, Kruusmaa M (2013) Flow-relative control of an underwater robot. Vu MT, Jeong SK, Choi HS, Oh JY, Ji DH (2018) Study on down-cutting ladder
Math Phys Engine Sci 469(2153):20120671 trencher of an underwater construction robot for seabed application.
Sankaranarayanan V, Mahindrakar AD, Banavar RN (2008) A switched controller Appl Ocean Res 71:90–104
for an underactuated underwater vehicle. Commun Nonlinear Sci Wang CC, Lu J, Ding XL, Jiang CX, Yang JY, Shen JH (2021) Design, modeling,
Numer Simul 13(10):2266–2278 control, and experiments for a fish-robot-based IoT platform to enable
Saxena M, Chauhan NR (2017) A review study on bio-inspired robotic fish. Int J smart ocean. IEEE Int Things J 8(11):9317–9329
of Mechan Robot Syst 4(1):1–23 Wang G, Song YJ, Tang WS, Xie GM, Li DJ (2019a) A numerical simulation analy-
Scaradozzi D, Palmieri G, Costa D, Pinelli A (2017) BCF swimming locomotion sis on bionic robot fish based on Computational Fluid Dynamics (CFD)
for autonomous underwater robots: a review and a novel solution to method. J Nanoelectron Optoelectron 14(3):400–407
improve control and efficiency. Ocean Eng 130:437–453 Wang LL, Wang HJ, Pan LX (2014a) Autonomous underwater vehicle motion
Serhat Y (2022) Development stages of a semi-autonomous underwater planning via sampling based model predictive control. Appl Mech
vehicle experiment platform. Int J Adv Robot Syst 19(3):15 Mater 670–671:1370–1377
Shen Q, Olsen Z, Stalbaum T, Trabia S, Lee J, Hunt R et al (2020) Basic design of Wang M, Dong HF, Li X, Zhang YL, Yu JZ (2019b) Control and optimization of
a biomimetic underwater soft robot with switchable swimming modes a bionic robotic fish through a combination of CPG model and PSO.
and programmable artificial muscles. Smart Mater Struct 29(3):1–24 Neurocomputing 337:144–152
Shen Q, Wang TM, Liang JH, Wen L (2013) Hydrodynamic performance of a Wang M, Zhang YL, Dong HF, Yu JZ (2020) Trajectory tracking control of
biomimetic robotic swimmer actuated by ionic polymer–metal com- a bionic robotic fish based on iterative learning. Sci China Inf Sci
posite. Smart Mater Struct 22(7):075035 63(7):170202
Shi L, Guo S, Asaka K (2010) A novel jellyfish-like biomimetic microrobot. 2010 Wang R, Wang S, Wang Y, Cai MX, Tan M (2019c) Vision-based autonomous
IEEE/ICME Int Conf Compl Med Engine 13(5):277–281 hovering for the biomimetic underwater robot-RobCutt-II. IEEE Trans
Industr Electr 66(11):8578–8588
Cui et al. Intelligent Marine Technology and Systems (2023) 1:7 Page 28 of 28

Wang SY, Han Y, Mao ST (2021) Innovation concept model and prototype vali- Yang YC, Ye XF, Guo SX (2007) A new type of jellyfish-like microrobot. 2007 IEEE
dation of robotic fish with a spatial oscillating rigid caudal fin. J Marine International Conference on Integration Technology, Shenzhen, China,
Sci Eng 9(4):435 20-24 March 2007, pp 673–678
Wang T, Wang Z, Zhang B (2021) Mechanism design and experiment of a Yang YH, Chu CZ, Jin H, Hu QQ, Xu M, Dong EB (2023) Design, modeling, and
bionic turtle dredging robot. Machines 9(5):86 control of an aurelia-inspired robot based on SMA artificial muscles.
Wang TM, Liang JH, Shen GX, Tan GK (2005) Stabilization based design and Biomimetics (Basel) 8(2):261
experimental research of a fish robot. IEEE/RSJ International Conference Yeom SW, Oh IK (2008) Fabrication and evaluation of biomimetic jellyfish
on Intelligent Robots and Systems, Edmonton, Canada, 2-6 August robot using IPMC. Symposium on Mining Smartness from Nature held
2005, pp 2065–2070 at the 3rd International Conference on Smart Materials, Structures and
Wang W, Bian XQ, Chang ZH (2002) 3D track-keeping method for autonomous Systems, Acireale, Italy, 8-13 June 2008, pp 171–176
underwater vehicle. J Mar Sci Appl 1(2):18–22 You SS, Lim TW, Jeong SK (2010) General path-following manoeuvres for an
Wang W, Wei XY, Meng K, Zhong LH, Wang Y, Yu XH (2017) Bio-tribology prop- underwater vehicle using robust control synthesis. Proc Inst Mech
erties of bionic carp scale morphology on Ti6A14V surface. IOP Conf Engine Part I: J Syst Control Eng 224(8):960–969
Ser: Mat Sci Engine 281:012060 Yu JZ, Wang C, Xie GM (2016) Coordination of multiple robotic fish with
Wang YD, Wang PC (2014) Design of a kind of minitype shallow water power applications to underwater robot competition. IEEE Trans Industr Electr
underwater vehicle. Appl Mechan Mat 644–650:674–677 63(2):1280–1288
Wang YW, Tan JB, Gu BT, Sang PF, Zhao DB (2014b) Design and modeling of a Yu JZ, Wen L, Ren ZY (2017) A survey on fabrication, control, and hydro-
biomimetic stingray-like robotic fish. Adv Mat Res 945–949:1473–1477 dynamic function of biomimetic robotic fish. Sci China Technol Sci
Wang YW, Tan JB, Zhao DB (2015) Erratum to: ’Design and experiment on a 60(9):1365–1380
biomimetic robotic fish inspired by freshwater stingray.’ J Bionic Eng Yu JZ, Wu ZX, Wang M, Tan M (2016) CPG network optimization for a
12(3):371 biomimetic robotic fish via PSO. IEEE Trans Neural Netw Learn Syst
Wang ZL, Wang YW, Li J, Hang GR (2009) A micro biomimetic manta ray robot 27(9):1962–1968
fish actuated by SMA. 2009 IEEE International Conference on Robotics Yue CF, Guo SX, Li MX, Li YX, Hirata H, Ishihara H (2015) Mechatronic system
and Biomimetics (ROBIO), Guilin, pp 1809–1803 and experiments of a spherical underwater robot: SUR-II. J Intell Robot
Wang ZW, Dong CL, Zhang ZM, Tian QH, Sun AQ, Yuan L et al (2022) Review Syst 80(2):325–340
of multi-fin propulsion and functional materials of underwater bionic Zhang SX, Yan WS, Xie GM (2017) Consensus-based leader-following formation
robotic fish. Proc Inst Mech Eng C J Mech Eng Sci 236(13):71 control for a group of semi-biomimetic robotic fishes. Int J Adv Robot
Watts CM, McGookin EW (2014) Surge performance of an underwater vehicle Syst 14(4):12
with a biomimetic tendon drive propulsion system. Proc Inst Mech Zhang Z, Yang T, Zhang T, Zhou F, Cen N, Li T et al (2021) Global vision-based
Engine, Part M: J Engine Maritime Environ 228(4):315–330 formation control of soft robotic fish swarm. Soft Robot 8(3):310–318
Wehner M, Truby RL, Fitzgerald DJ, Mosadegh B, Whitesides GM, Lewis JA et al Zhao W, Yu J, Fang Y, Wang L (2006) Development of multi-mode biomimetic
(2016) An integrated design and fabrication strategy for entirely soft, robotic fish based on central pattern generator. 2006 IEEE/RSJ Interna-
autonomous robots. Nature 536(7617):451–455 tional Conference on Intelligent Robots and Systems, Beijing, China,
Wu RX, Du G, Liu Z, Zhang DX, Yu YJ (2021a) Design of bionic robot fish pro- 9-13 October 2006, pp 3891–3896
pelled by two joint caudal fin. J Phys: Conf Ser, 1982:012056 Zheng XW, Wang W, Xiong ML, Xie GM (2020) Online state estimation of a fin-
Wu YH, Liu JC, Wei YG, An D, Duan YH, Li WS et al (2021b) Intelligent control actuated underwater robot using artificial lateral line system. IEEE Trans
method of underwater inspection robot in netcage. Aquac Res Robot 36(2):472–487
53(5):1928–1938 Zhong Y, Song JL, Yu HY, Du RX (2018) Toward a transform method from light-
Wu YH, Ta XX, Xiao RC, Wei YG, An D, Li DL (2019) Survey of underwater robot hill fish swimming model to biomimetic robot fish. IEEE Robot Automat
positioning navigation. Appl Ocean Res 90:101845 Lett 3(3):2632–2639
Wu ZX, Yu JZ, Su ZS, Tan M (2013) Control and implementation of S-start Zhou C, Cao ZQ, Wang S, Tan M (2008) Study on the pitching and depth con-
for a multijoint biomimetic robotic fish. Acta Automatica Sinica trol of biomimetic robot fish: study on the pitching and depth control
39(11):1914–1922 of biomimetic robot fish. Acta Automatica Sinica 34(9):1215–1218
Wynn RB, Huvenne VAI, Le Bas TP, Murton BJ, Connelly DP, Bett BJ et al (2014) Zhou J, Huang C, Liu B, Wang TT (2014) A control system design of under-
Autonomous Underwater Vehicles (AUVs): their past, present and future water robot for detecting underwater structures. Appl Mech Mater
contributions to the advancement of marine geoscience. Mar Geol 536–537:1105–1109
352:451–468 Zhou L, Dong EB, Hang H, Kong XW, Yang J (2017) Design and research of a
Xia QC, Li H, Song N, Wu ZL, Wang X, Sun X et al (2023) Research on flexible novel caudal-fin propulsion mechanism with two degrees of freedom.
collapsible fluid-driven bionic robotic fish. Ocean Eng 276(8):1–10 IEEE International Conference on Real-time Computing and Robotics
Xie FR, Li Z, Ding Y, Zhong Y, Du RX (2020) An experimental study on the fish (RCAR), Okinawa, Japan, 14-18 July 2017, pp 454–458
body flapping patterns by using a biomimetic robot fish. IEEE Robot Zhou ZY, Jiang YQ, Li Y, Jian C, Sun YY (2018) A single acoustic beacon-based
Automat Lett 5(1):64–71 positioning method for underwater mobile recovery of an AUV. Int J
Xie FR, Zuo QY, Chen QL, Fang HT, He K, Du RX et al (2021) Designs of the Adv Robot Syst 15(5):1729881418801739
biomimetic robotic fishes performing Body and/or Caudal Fin (BCF) Zhou ZY, Liu JC, Pan J, Yu JZ (2023) Proactivity of fish and leadership of self-
swimming locomotion: a review. J Intell Robot Syst 102(1):13 propelled robotic fish during interaction. Bioinspir Biomim 18(3):1748
Xie O, Zhu QX, Shen L, Ren K (2018) Kinematic study on a self-propelled bionic Zhu KB (2018) Tracking and positioning technology of chemical plumes by
underwater robot with undulation and jet propulsion modes. Robotica underwater robots based on source distribution model. Chem Engine
36(11):1613–1626 Trans (CET Journal) 71(1):451–456
Yan CY, Zhang XQ, Ji ZY, Wang XL, Zhou F (2021) 3D-Printed electromag-
netic actuator for bionic swimming robot. J Mater Eng Perform
30(9):6579–6587 Publisher’s Note
Yan Q, Han Z, Zhang SW, Yang J (2008) Parametric research of experiments on Springer Nature remains neutral with regard to jurisdictional claims in pub-
a carangiform robotic fish. J Bionic Eng 5(2):95–101 lished maps and institutional affiliations.
Yang H, Ma J (2010) Nonlinear control for autonomous underwater glider
motion based on inverse system method. J Shanghai Jiaotong Univ
(Science) 15(6):713–718
Yang J, Feng JF, Qi D, Li YL (2016) Longitudinal motion control of underwa-
ter vehicle based on fast smooth second order sliding mode. Optik
127(20):9118–9130