An autonomous quadrotor delivery system

to help oldsters & disabled persons

And facilitating services to government,
companies and citizens

Menna Ahmed Megahed

Second Year of Preparatory

Zahraa Islamic Language School

1
 Contents
Title Page
Cover page 1
Contents 2
Objective of the project 3
Problem formulation 3
Idea description 4
Introduction 4
Previous work 5
Specification and setup 6
Scenario 7
Software diagrams 9
Agent specification 11
Required hardware 11
Proposed system 13
Tools 14
Evaluation and Results 14
Conclusion 16
Future work 16
References 19
2
1- Objective of the project:

The aim of this project is to propose an Autonomous Quadrotor Delivery

System, designed for disabled and seniors to help them get their work done. It
also addresses the problem of crowd due to our drone would make its work
quickly.

2- Problem formulation:

Statistics about disabled people say that about 2 of each 10 adults live with a
disability. About 60 percent of those, their disability is in moving. Sothey have
many problems when they want anything outdoor and also lowers their quality of
life since they must rely on some sort of aid to get around.

Seniors, the older population, persons 60 years or older are increasing around the
world. Statistics about them say that they are about 15-20 percent of people and
they are increasing. It’s expected to be 30 percentby 2050. These people usually
suffer from geriatrics and also may need some sort of need to get anything
outdoor or to get around.

In Egypt and in many cities around the world there’re a problem of crowdthat
delay delivery of many things for us for many times, if that delay wasshortened
or removed there would save many times. Crowd also may result in dangerous
results in cases like the delay of delivering medicine in time.
You could apply the dangerous consequences of delay in delivery in suchcrowd
cities on other cases.

A good solution will be a device that could replace traditional delivery systems with
others that provide delivery quickly so that it could save timeand effort. That
device could be owned either by individuals or companies or governments. For the
purpose of this project, we streamlined the problem defined above to help
disabled and seniors get what they want atthe place they want and to replace
traditional delivery system with another that would save time so that we get rid of
the problem of crowd and its consequences on delivery systems.

3
3- Idea description:

For government or companies

1. Customer or citizen would log into the site and then request the service for
example: request for identity card.
2. An application would give the citizen's location to the drone.
3. Drone would reach the place where the customer would take the identity
card.
4. Drone would return to the government.
For individuals
Customer can order any service by send HomingDrone to place and then it back
with order example: request for a meal.
1. An application would give the service location to the drone.
2. Drone would reach the restaurant where the employees would put the
meal in it. employees communicate with customer by screen and camera in
the Drone so customer could see employees and employees could see him
and when he agrees the money customer withdraw money from
HomingDrone via the application of HomingDrone (i.e. our HomingDrone
have a machine that looks like ATM but customer only who can withdraw).
3. Drone would return to its user.
4. During that the owner could monitor it by a provided application.

4- INTRODUCTION

A quadrotor is a rotary-wing Unmanned Aerial Vehicle (UAV) that has four

rotors. To permit a quadrotor to navigate autonomously in outdoor environments is
very difficult. A quadrotor is not easy to control, since it is considered as an under
actuated system for it has six Degrees of Freedom (DOF) but only four actuators.
Fusing inertial measurements and features in images captured from camera is
known as a vision-aided inertial navigation.
The goal of this project is to present the specification and the developmentof an
autonomous quadrotor delivery system that fuses data from camera and IMU
(Inertial Measurement Unit) to allow for in-time delivery of products.

4
5- PREVIOUS WORK

There are a wide range of current and potential civil applications. That includes: Load-
Delivery, Hostile Load-Delivery Recently, Passenger Transport, Journalism, Law
Enforcement, Veronalism, Community Policing, Hobby and Entertainment Uses.

These figures represent some examples on UAV's

5
 Many big firms are eager to utilize flying robots in delivery.

Amazon: Amazon Prime Air is a conceptual drone-based delivery system

currently in development by Amazon company which is the biggest company in
the world
https://www.youtube.com/watch?v=Le46ERPMlWU
Domino's: Domino's has tested out a prototype for a flying drone that can deliver
pizzas. The device carried two large pepperoni pizzas and conducted a flight.
https://www.youtube.com/watch?v=-CYT4PFV_Hs
Francesco's Pizzeria: Francesco's Pizzeria provides a way for delivering pizza
using flying robots.
https://www.youtube.com/watch?v=0if2PM6OBrI

6- SPECIFICATION AND SETUP

Drone introduces public services that could be used either by the government,
companies or by individuals. It enables a customer or a citizen to request the
required service from the site of the company or the government and delivered to
the customer or citizen in his/her location using the drone. It also could be owned
and used by individuals to meet their requirements. The drone has an application
that prevent any errors from occurrence. It also provide the owners to monitor it
using a site.
a. The importance of this idea

The goal of the idea is to save effort and time so it's useful for oldsters who aren't
able to move and get their work done. It's also useful for saving time in such
countries that have a problem of crowd, it's also can be useful in emergency
states.

b. Video description
A video description of our research
https://www.youtube.com/watch?v=oARfFS34z5E&ab_channel=MennaAhmed

6
7- Scenario
First scenario
INITIAL ASSUMPTION:
Person need a service from any company or government, he requests service from website and
employee do service and send service (product) to customer by drone
NORMAL:
employee receive request of customer he do service, then ask drone to carry product by using
put method and get location of customer through GPS (get place function) and put location to
system and use fly method to send drone to customer with product and use home method to
return drone to company
He can monitor drone during the way by monitor method

WHAT CAN GO WRONG:

Drone may be need maintenance or anything goes wrong during the way to destination.
Employee may not able to access the system if he forgets user name or password

OTHER ACTIVITIES:
Employee is the only person that control drone

SYSTEM STATE ON COMPLETION:

Drone return to government or company and new customer is added in database

State diagram for user scenario

7
Second scenario
INITIAL ASSUMPTION:
Disabled or senior need any service or need to eat they can send drone to any place to do the
order and can return to home

NORMAL:
disabled or senior get location of the restaurant by get place method then put location to system
and use fly method to send drone to the restaurant and he can monitor the drone during the way
and he can communicate with employee by screen and camera by talk and monitor method and
he can pay by PayPal by pay method and then return drone to home by home method

WHAT CAN GO WRONG:

Drone may be need maintenance or anything goes wrong during the way to destination.
Disabled or senior may not able to access the system if he forgets user name or password
Disabled or senior account don't have enough money

OTHER ACTIVITIES:
Disabled or senior is the only person that control drone

SYSTEM STATE ON COMPLETION:

Drone return to home and new place is added in database.

State diagram for Employee scenario

8
8- Software Diagrams
8.1 context diagram

8.2 block diagram

9
8.3 Business model canvas

8.4 entity relationship diagram (ERD)

10
9- Agent specification
One of high-level specifications is PEAS (Performance measure, Environment,
Actuators, and Sensors) as an agent which is shown in table I.

TABLE I: PEAS specification

Performance Measure Reaching destination rapidly

Collaboration among quadrotors

Future Performance Measure
into a swarm

Environment Buildings, suspicious Objects

Actuators 4 rotors
Sensors Camera and IMU

10- Required Hardware:

We are acquiring an AR.Drone ParrotDrone (http://ardrone2.parrot.com/)
With the following specifications
Model Parrot AR drone 2.0
Battery 1500mAh LiPo
Takeoff weight 420g
The motor has a power of 65Watts and
Motors completes 28,000 revolutions per minute(RPM) when
hovering.
This kit includes 4 propellers (2 "C" and 2 "A"),
Propeller
and 8 circlips
4 brushless in runner motors. 14.5W 28,500
Motors
RMP
Camera HD Camera. 1080p 30FPS
Communication Wi-Fi and satellite

Ultrasound Ultrasound sensors for ground altitude measurement

60 FPS vertical 60 FPS vertical QVGA camera for ground speed

QVGA camera measurement
Storage 16GB DDR4 RAM at 200MHz
11
The following figure represent the Parrot AR drone 2.0

Overall Requirements:
Drone must have an IMU, slam, flight controller, two digital cameras,
microcontroller, and a data link module
Drone must be autonomous
Manual override must be available at all times during flight
Drone must be able to adjust its flight to avoid obstacles
Drone shall generate its own waypoints to capture images of the entire user
defined area
Drone must be able to combine all of the images acquired into one high resolution
image
Drone shall alert the user that the survey is complete via email, using the data link
module
Drone shall complete task within 30 minutes
Drone must have station to charge its battery
Drone should carry weight closer to 20 kg
Plan must build local map and determine location
Airplane Requirements:
o The drone must be electrically powered
o The drone must be hand launched
o The battery must last long enough for the entire task to be done in one charge(~20
minutes)
o The plane must be able to operate while holding all components
Camera Requirements:
➢ One front facing camera shall be of lower resolution to be used for obstacle
detection
➢ Ground facing camera must be high resolution with the ability to capture near
infrared pictures.
➢ All images captured with the ground facing camera must be tagged with GPS.

12
Delivery system Requirements:
• website for customer to request his order
• employee (customer) login system
• employee do the order and put it in the drone
• employee determine the location of the customer and send the drone to this location
• employee can monitor the drone during the way to destination through monitor
• when it arrives to its destination employee use home method to return to company
or government
• customer send drone to the place where he requests his order
• he can communicate with employee through camera and screen
• Customer pay through PayPal.
• Customer use home method to return drone.
• drone use slam to make localization and build mapping.

11- PROPOSED SYSTEM

Figure 1 represents the diagram of control unit of most autonomous robots. The
quadrotor has to solve the simultaneous localization and mapping (SLAM) problem in
order to be autonomous So it has to specify its location within an altering map that it
constructs of the environment based on data fusion from many data sources
Developments use visual or sonar navigation. The frequent utilized filter is Kalman
Filter, which is similar to a real-valued Hidden MarkovModel with Gaussian
probability noise.

Fig 1 control unit of autonomous robots

13
12- Tools

Modern simulators tend to provide the following features:

• Fast robot prototyping

• Using the own simulator as creation tool (Virtual Robot Experimentation

Platform, Webots, R-Station, Marilou).

• Using external tools.

• Physics engines for realistic movements. Most simulators use ODE (Gazebo,
LpzRobots, Marilou, Webots) or PhysX (Microsoft Robotics Studio).

• Realistic 3d rendering. Standard 3d modeling tools or third-party tools can be

used to build the environments.

• Dynamic robot bodies with scripting. C, C++, c#, Perl, Python, Java, URBI,
MATLAB languages used by Webots, Python used by Gazebo.

13- Evaluation and Results

The next figure is the result of canny edge detection

14
The next figure is the result of Autonomous “SLAM”

15
14- Conclusion

The goal of this project is to present the specification and the development of
an autonomous quadrotor delivery system that fuses data from camera and
IMU(Inertial Measurement Unit) to allow for in-time delivery of products. The results of
simulation and experiments are presented, the experiments on the collaboration
among quadrotors into a swarm are in progress. Future work include using solar
energy for charging and make a collaboration among quadrotors as a swarm.

15- Future work

Swarm
The future work includes the collaboration among quadrotors as a swarm are in
progress like in the following figure

Swarm collaboration

16
 Map Implementation
The future work also includes applying our project on the map of the Smart village
which is presented in the next figure

Smart Village map

SLAM Future Work

The Future Work in slam an extension of the work done in this project might be to
combine the three algorithms described in this thesis in a way that yields more robust
performance. There is no reason a robot would never travel from an environment with
many available features to one without features. Additionally, a robot might be able to
observe both areas with features and areas without from a single location. It seems
that if a robot were able to use both a laser range finder to employ DP-SLAM and a
camera to employ FastSLAM then it would be better prepared to handle more diverse
environments. SLAM methods combining data from camera sensors and laser range
finders were considered before FastSLAM or DP-SLAM were developed and this
yielded greater accuracy than other contemporary methods. It may be that combining
the new algorithms we have discussed could yield even better results.

17
 Business
The key activities will you undertake in this project in (+3 years):

I plan to spread I also plan to

our technology make
I also plan partnerships
in Egypt so that when our
people get trust with
product is used organizations
of the product in a wide range
and to get that help
to use them as oldesters and
familiar with it a folk so that
so the product disables
each drone is
could be used monitoring we also plan to
any where like other drones. export our
streets securely drone to
without any abroad
fear

18
16-References
1. Philip H.S. Torr and Andrew Zisserman: Feature Based Methods for Structure and
Motion Estimation, ICCV Workshop on Vision Algorithms, pages 278-294, 2023

2. Michal Irani and P. Anandan: About Direct Methods, ICCV Workshop on Vision
Algorithms, pages 267-277, 1999.

3. Feature Detectors - Sobel Edge Detector Moeslund, T. (2020, March 23). Canny
Edge Detection. Retrieved December 3, 2020

4. Glyn W. Humphreys and Vicki Bruce (1989). Visual Cognition. Psychology

Press.ISBN 0-86377-124-6.

5. B. Glocker, N. Komodakis, G. Tziritas, N. Navab& N. Paragios (2021). Dense

Image Registration through MRFs and Efficient Linear Programming (PDF).
Medical Image Analysis Journal.

6. http://vision.middlebury.edu/flow/

7. Jun Li, Yuntang Li,” Dynamic Analysis and PID Control for a Quadrotor”,
Proceedings of the 2021 IEEE International Conference on Mechatronics and
Automation, August 7 - 10, Beijing, China.

8. Pieter J. Mosterman, David Escobar Sanabria, Enes Bilgin, Kun Zhang, Justyna
Zander, Heterogeneous Fleet of Vehicles for Automated Humanitarian Missions,
Computing in Science and Engineering, vol. 16, no. 3, pp. 90-95, May-June, 2022.

9. Hartman, K. Landis, M. Mehrer, S. Moreno, J. Kim “Quadcopter Dynamic Modeling

and Simulation”2022.

10. S. Bouabdallah, P. Murrieri, and R. Siegwart, Design and Control of an Indoor

Micro Quadrotor. In Proc. ICRA, New Orleans, Louisianna, 2004

19
20