Design and Implementation of an Autonomous Water Surface

Cleaning Robot
1Mina Yousef Takawy a
2Ahmed Gamal Ezz-Eldin b, 3Amgad M. Bayoumy Alyc
a
B.Sc. Student, Faculty of Engineering, MSA University, 26 July Mehwar Road
intersection with Wahat Road, 6th October City, 12451, Egypt.
b
B.Sc. Student, Faculty of Engineering, MSA University, 26 July Mehwar Road
intersection with Wahat Road, 6th October City, 12451, Egypt.
c
Associate Professor, Faculty of Engineering, MSA University, 26 July Mehwar
Road intersection with Wahat Road, 6th October City, 12451, Egypt.
E-mail: a [email protected], b [email protected] ,
[email protected],eg

Abstract

Water is a vital resource, but industrialization has caused water pollution to

become a significant issue. An autonomous water surface-cleaning robot is a
machine that is designed to navigate and clean the surface of a body of water
without human intervention. The robot is equipped with sensors and algorithms
that allow it to detect and remove debris, such as leaves and twigs, from the water's
surface. In addition to removing visible debris. one of the main advantages of using
an autonomous water surface cleaning robot is the ability to continuously clean the
water without the need for human intervention. This allows for a more efficient
and effective cleaning process, as the robot can operate around the clock if needed.
Additionally, the use of a robot can help to reduce the risk of injury or accidents
that may be associated with manual cleaning methods, such as using a net to
remove debris. A fixed arm is included for collecting and depositing waste into a
basket on the hull, and electronic circuits and motors are protected inside the hull.
ENGINEERING JOURNAL Volume 2 Issue 2
Received Date January 2023
Accepted Date March 2023
Published Date March 2023
DOI: 10.21608/MSAENG.2023.291921
 DOI:10.4186/ej.2009.VOL.ISSUE.pp

The robot can operate autonomously, including detecting and avoiding obstacles
and cleaning the water's surface. It also has a docking station that alerts the robot
when it has reached its maximum garbage capacity or when the battery is low. The
robot includes modules for movement, positioning, obstacle avoidance,
communication, and power management, as well as a system for cleaning the
water's surface.
Keywords: Unmanned Systems, Boat, Control loops, Navigation and Control,
Robots

1. Introduction

The proposed project aims to design and implement a robot capable of cleaning the water
surface of floating garbage in the Nile River. This is necessary due to the negative impact of
water pollution on the availability of clean drinking water and the health of aquatic organisms.
The current methods, such as semimanual vessels, have limitations in their ability to
accurately collect only garbage and avoid causing secondary pollution. The proposed solution
is to design a boat with an autonomously controlled system that utilizes a camera, sensors,
and object detection system to collect the garbage while minimizing the impact on the
environment. The project will involve reviewing and summarizing previous work, developing
a preliminary design, conducting motion simulation, and implementing and testing the robot.
The goal is to improve the water quality and preserve the aquatic ecosystem of the Nile River.

2. Literature Review on Water Surface Cleaning Robot

2.1. Autonomously Controlled

An intelligent robot system for cleaning plastic trash from the surface of bodies of water is
being developed. The robot can detect, track, steer, and collect debris. There are challenges
in implementing these capabilities, including accurately and efficiently recognizing garbage,
maintaining stability during steering based on vision, and successfully capturing floating
debris despite disruptions on the water's surface. The robot is composed of a vision module,
motion control module, and grasping module, and can complete three tasks in sequence:
cruising and detecting, tracking, and steering, and grasping and collecting. In the first task,
the robot navigates a predetermined path on the water's surface and uses its vision module to
search for trash. When an object is identified for removal, the second task begins. The vision
module tracks the target and calculates the robot's position in relation to it, while the motion
control module uses this position data to adjust the yaw angle error for a precise approach
angle. The grasping module then determines when to initiate the final task, during which the
manipulator grabs and collects the object. The technology is based on an electric underwater
vehicle that can travel on the water's surface while carrying a load. The sliding-mode
controller approach has been effectively applied as the motion control module due to its
insensitivity to model parameters [7].

2 MSA ENGINEERING
JOURNAL
Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339
(https://msaeng.journals.ekb.eg//)
 Recycling the garbage collected by the robot not only helps to clean the water body, but
also reduces the carbon footprint of creating new products. Pollutants found in water bodies
can be recovered through this process. By removing pollution and monitoring water quality,
the robot aims to clean the body of water. While there are several methods for cleaning
pollution in the open ocean, there are fewer options for cleaning coastal or rocky water bodies.
The robot is effective in cleaning these types of water bodies, especially when large-scale
approaches are not feasible. The robot can be used to gather data on pollution and water
quality, which can be used to draft local regulations that benefit society. It can be controlled
remotely or through an app, and has various features such as automatic cleaning, route
planning, edge cleaning, and automatic return. The robot, developed by Ocean Alpha, can
identify and avoid obstacles in real-time using millimetres wave radar, lidar, and an intelligent
sensing unit. It can also recognize target garbage and adjust its cleaning strategy through
ongoing machine learning. The platform can be modified to perform additional tasks such as
hydrographic survey, reagent distribution, water quality monitoring, and water plant mowing
[3].

The proposed detecting approach allows the capturing robot to effectively clear floating
trash in the field. Experimental results show that the modified you only look once v3
(YOLOv3) object detection algorithm performs better than other object identification
algorithms in terms of both detection speed and accuracy. This is important for the
autonomous and intelligent high-speed recognition and grabbing of moving objects in
complex aquatic environments. During clean-up operations, the robot has been able to detect
and capture a variety of floating trash such as plastic bottles, plastic bags, and Styrofoam. The
YOLOv3 network is used for detection, which is based on the first 52 layers of the DarkNet-
53 network and can extract features from images and recognize and classify multi-class
objects from 3 scales and 9 ranges [9].

2.2. Remotely Controlled

The design for a waste collection system that effectively removes trash and debris, such as
litter, tires, and logs, from rivers, channels, and lakes is part of a larger system. The system
utilizes IoT technology to monitor and manage the waste collection process. The vessel, which
is designed to remove toxins from waterways and work in areas beyond the shoreline, aims
to provide more options for cleaning up the aquatic environment. The river-cleaning robot is
made up of three components. The first section deals with the input source, including a mobile
phone app with a magnetic switch and a solar panel. The second section focuses on processor
development, using Arduino IDE software to code the microcontroller processor with
NODEMCU V2. The third section addresses the output source, which combines mechanical
components such as DC motors [6].
A motor-driven collecting-arm system has been developed for the efficient removal and
redistribution of waste into a rectangular basket on the hull. The prototype's hull is made of
Styrofoam and coated in waterproof epoxy, supported by an aluminium wire frame. The
propulsion system, based on a differential drive mechanism, provides strong thrust and the
ability to turn 360 degrees. Electronic circuitry and motors are installed in the hull to keep
them dry, and the robot can be controlled using a remote with Xbee Pro wireless modules.
Testing showed that the robot can effectively collect and carry up to 16 kg of rubbish at a time
[14].

3 MSA ENGINEERING JOURNAL

Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)
 This study presents a design plan for a water surface collection robot intended for use in
deep-sea cage culture. The robot is equipped with a creatively designed forearm drainage and
retraction mechanism and has been tested in a prototype and experimental studies on sea
surface collection. These studies have demonstrated the viability and rationality of the
designed water surface collection robot. The robot is remotely controlled to collect dead fish,
which not only improves the efficiency of collecting and cleaning dead fish in the cage culture
area but also reduces the workload and potential safety risks for the cage culture crew. The
experimental results show that the robot is a simple and effective method for collecting and
cleaning dead fish on the water's surface in the deep and open sea cage culture region, offering
a better accuracy trade-off compared to other marine waste collecting techniques [10].

According to a different approach, the bot will search for waste patches along the road and
collect the rubbish, clearing the waterways in the process. In this method, if someone is found
poisoning the water sources, a monitoring system alerts the authorities. A more sustainable
system has been made using solar energy. The use of computer vision algorithms allows for
the detection of trash on the water's surface. This trash is collected by the bot, who then dumps
it in a designated location. In addition to cleaning the water bodies, preventive measures like
virtual fencing have been utilized to alert the authorities if someone tries to contaminate the
water premises. The movement of the bot and shore surveillance are both monitored by web
applications and mobile apps, respectively. Both preventive and curative steps are included
in this comprehensive solution for water maintenance. The surveillance system will activate
a camera at the shore to look for anyone attempting to contaminate the waterways by
employing image processing methods. Image processing algorithms are used for person
detection. Based on the discovery, an alarm will be sent via the app to the authorities so they
can take the appropriate action. Using deep learning and OpenCV models, the image
processing technique is utilized to recognize human activity along the coast and classify it as
either polluting or not polluting water bodies. To use the app to access the live video feed
produced by the cameras (both the one mounted on the boat and the one used for offshore
monitoring); they use the custom TCP tunnel service offered by Remote.it. Remote. it
establishes a tunnel connection between TCP port 8081 of the Raspberry Pi (which runs
motion) and the end device which allows the video stream to be accessed remotely [2].

3. Simulation

3.1. Vehicle Dynamic Model

The behaviour of ships can be explained by utilizing highly nonlinear equations based on
Newton's second law (Gierusz,2016).

𝑚(𝑢 + 𝑞𝑤 − 𝑟𝑣) = 𝑋𝑡𝑜𝑡 Eq. (1)

𝑚(𝑣 + 𝑟𝑢 − 𝑝𝑤) = 𝑌𝑡𝑜𝑡 Eq. (2)

𝑚(𝑤 + 𝑝𝑣 − 𝑞𝑢) = 𝑍𝑡𝑜𝑡 Eq. (3)

4 MSA ENGINEERING JOURNAL

Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)
 𝐼𝑥𝑝 + (𝐼𝑧 − 𝐼𝑦)𝑞𝑟 = 𝐾𝑡𝑜𝑡 Eq. (4)

𝐼𝑦𝑞 + (𝐼𝑥 − 𝐼𝑧)𝑟𝑝 = 𝑀𝑡𝑜𝑡 Eq. (5)

𝐼𝑧𝑟 + (𝐼𝑦 − 𝐼𝑥)𝑝𝑞 = 𝑁𝑡𝑜𝑡 Eq. (6)

where:
The mass of the ship is represented by m, while the moments of inertia around X, Y,
and Z axes are represented by 𝐼𝑥 , 𝐼𝑦 , and 𝐼𝑧 respectively. The total forces acting along the X,
Y, and Z axes are represented by 𝑋𝑡𝑜𝑡 , 𝑌𝑡𝑜𝑡 , and 𝑍𝑡𝑜𝑡 , respectively. The total moments acting
on the hull are represented by 𝐾𝑡𝑜𝑡 , 𝑀𝑡𝑜𝑡 , and 𝑁𝑡𝑜𝑡 .
In many control scenarios such as stabilization of heading, speed, and trajectory, it is
typical to disregard the pitch, roll and heave motion by assuming that p = q = w = 0. This
simplifies the 6-degree of freedom model of the ship to a 3-degree of freedom model as shown
in Fig.1.

𝑚(𝑢 − 𝑟𝑣) = 𝑋tot [N] Eq. (7)

𝑚(𝑣 + 𝑟𝑢) = 𝑌tot [N] Eq. (8)
𝐼𝑧𝑟 = 𝑁tot [Nm] Eq. (9)

Fig. 1 the Earth-fixed reference frame X0-Y0-Z0 and the ship-fixed.

(Gierusz,2016)

5 MSA ENGINEERING JOURNAL

Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)
 3.2. Vehicle Dynamic Model Simulation

Fig. 2 Dynamic Model Simulation

The vehicle dynamic is implemented on MATLAB/Simulink as mentioned in Fig.2 above.

The two input forces are F1 which is a force acting from one of the propellers and F2 acting
from another propeller.

4. Hardware design
After studying the environment in which the robot will operate, the proposed autonomous
system will face various challenges. For example, the Nile River is a dynamic environment
that is subject to fluctuating water levels and currents. The robot will need to be able to
navigate these changing conditions while maintaining its stability and collecting the garbage.
Additionally, the robot will need to be able to withstand the harsh weather conditions in the
Nile River, such as high temperatures and intense sunlight. Furthermore, the robot will need
to be able to detect and collect a variety of different types of garbage, including plastics and
other debris that may be floating on the water's surface.
However, the project also has several constraints that must be taken into consideration.
First, the robot must be safely isolated to avoid short circuits and preserve its components
from water. Second, the robot speed is capped at 0.7 m/s to ensure its stability and safety.
Third, the batteries should help the robot to work for more than 2 hours to make sure that the
robot does not as well as make sure that the robot can reach its recharging destination safely.
Fourth, the total voltage supplied to the robot is 12 V. Fifth, the communication range is 2m
Obstacle Avoidance. Sixth, the single cleaning capacity is 8 kg. Lastly, the robot Material is
PVC.
Overall, the environment in which the robot will operate is complex and challenging, but
with the appropriate design and technology, it has the potential to be a powerful tool for
cleaning the Nile River and preserving the aquatic ecosystem while considering the project
constraints.

6 MSA ENGINEERING JOURNAL

Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)
 After studying the environment in which the robot will operate and taking all the
constraints into consideration, we have developed a 3D CAD model of the robot using CAD
software as shown in Fig.3.
Table 1.

“Robot Specifications”
Design parameter Value
Dimensions (overall) 80x60x40cm
Net weight (only robot) 40kg
Maximum weight (with load) 48kg
Average battery while cleaning 2 hours
Propeller type
three blades
Control system Autonomous

Fig. 3 Robot design(mm)

7 MSA ENGINEERING JOURNAL

Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)
 The proposed robot design incorporates various components to deliver its
functionality, including actuators, sensors, and additional crucial elements. The actuators of
the robot include brushless DC motors, which provide a high-efficiency and reliable means
of propulsion. The brushless DC motor is known for its high efficiency and precision which
is crucial for the robot to navigate in the dynamic conditions of the Nile River.

Sensors play a crucial role in the robot's ability to navigate and collect garbage. The
robot is equipped with an IMU (Inertial Measurement Unit) sensor which provides
information about the robot's orientation and motion. This sensor is crucial for the robot to
maintain its stability and navigate the changing conditions of the Nile River.

Additional crucial elements include a GPS module, which provides the robot with its
location and allows it to navigate to specific areas of the Nile River. The robot is also equipped
with a Li-ion battery, which provides a reliable power source and ensures that the robot can
operate for more than 2 hours. Additionally, the robot is equipped with a Raspberry pi Camera
module, which allows the robot to detect and collect garbage and a relay module that helps
the robot to avoid obstacles. All these components are integrated to work together as shown
in Fig.4 to deliver the robot's functionality and achieve its goal of cleaning the Nile River and
preserving the aquatic ecosystem.

Fig. 4 circuit design

8 MSA ENGINEERING JOURNAL

Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)
 The nature of the environment where the robot will operate. Also, the robot CAD model
with their dimensions and this done using 3D CAD tools to draw the robot and simulate it, as
well as a detailed examination of the process for selecting motors for the robot and comparing
them to other options. The work environment and its characteristics were examined
thoroughly, then the selection criteria for the appropriate motors were discussed and different
types of motors were evaluated and compared. Also showing the criteria of selecting the
appropriate microcontroller, sensor, and other essential components and showing the
component's wiring diagram.

5. Results
The simulation conducted in MATLAB provides insight into the dynamics and
performance of a particular system under various conditions. The results show the system's
behaviour and performance under different scenarios.

5.1. General case

The robot's speed is moving in a straight line, as indicated by the absence of motion in
the Y direction and the Fig.5 shows that there are no steering angles present in this case.

Fig.5 General case on XY graph

5.2. First Case

The results of the first case show the situation when the speed of motor two is increased
with respect to motor one, the robot initiates movement in an anti-clockwise direction. This
movement continues until the robot reaches the necessary angle to effectively collect the
garbage as identified by the camera. In addition, the Fig.6 below represents the X Y graph
which describes the critical movement of the robot with this steering angle.
.

9 MSA ENGINEERING JOURNAL

Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)
 Fig.6 First case on XY graph

5.3. Second case

In the second case, when changing the speed of motor two less than motor one. The robot
starts moving clockwise until reaches the required angle to collect the garbage that was
captured by the camera. In addition, the Fig.7 below represents the X Y graph which describes
the critical movement of the robot with this steering angle which indicate that as the speed of
the robot changes than the first case the robot will rotate in the opposite direction of the first
case.

10 MSA ENGINEERING JOURNAL

Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)
 Fig.7 Second case on XY graph

6. Prototyping

To prove the concept of design we have manufactured a scaled prototype as shown in fig.8
to verify the concept.

Fig.8 Prototype

7. Conclusion
In conclusion, the project presents the design and implementation of an autonomous water
surface-cleaning robot that can collect floating garbage from water surfaces. The robot's
design ensures stability, navigation, and payload capacity, making it an efficient tool for
11 MSA ENGINEERING JOURNAL
Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)
cleaning the Nile River and preserving the aquatic ecosystem. Additionally, the robot uses
changes in motor speeds to navigate and collect garbage efficiently. It adjusts its movement
in response to the presence of garbage and the required steering angle, as observed in figures,
effectively navigating, and collecting garbage.

8. References
1. Adarsh, J. K., Anush, O. S., Shrivarshan, R., Mithulesh Krishnaan, S., Akash, J. K.,
Arul, R. and Angalaeswari, S. (2021) Ocean Surface Cleaning Autonomous Robot
(OSCAR) using Object Classification Technique and Path Planning Algorithm, In
Journal of Physics: Conference Series, IOP Publishing Ltd.

2. Amarnath, M., Bhuvanashankar, P., Ganesh, V. K. and Bindu, S. B. (2021)

WATERBIN-A Remote Controlled Water Surface Cleaning RoboT, International
Research Journal of Engineering and Technology, [online] Available at:
www.irjet.net.

3. Anon (2018) Piranha: An Autonomous Water Surface Robot, anon (n.d.) Water
Surface Cleaning Robot MC120, [online] Available at: www.oceanalpha.com.

4. Chang, H. C., Hsu, Y. L., Hung, S. S., Ou, G. R., Wu, J. R. and Hsu, C. (2021)
Autonomous water quality monitoring and water surface cleaning for unmanned
surface vehicle, Sensors (Switzerland), MDPI AG, 21(4), pp. 1–21.

5. Gao, X. and Fu, X. (2020) Miniature Water Surface Garbage Cleaning Robot, In
Proceedings - 2020 International Conference on Computer Engineering and
Application, ICCEA 2020, Institute of Electrical and Electronics Engineers Inc., pp.
806–810.

6. IEEE Control Systems Society. Chapter Malaysia and Institute of Electrical and
Electronics Engineers (n.d.) Proceedings, 2020 16th IEEE International Colloquium
on Signal Processing & its Application (CSPA 2020) : 28th-29th February 2020 :
conference venue, Hotel Langkawi, Lot 1852 Jalan Penarak, Kuah 07000 Langkawi,
Kedah, Malaysia,.

7. Kong, S., Tian, M., Qiu, C., Wu, Z. and Yu, J. (2021) IWSCR: An Intelligent Water
Surface Cleaner Robot for Collecting Floating Garbage, IEEE Transactions on
Systems, Man, and Cybernetics: Systems, Institute of Electrical and Electronics
Engineers Inc., 51(10), pp. 6358–6368.

8. Kumar Dutta, P., Nag, S., Chakraborty, M. and Nandy, S. (2021) An Automated
Surface Water Cleaning Robot for Ponds and Water Reservoirs,.

9. Li, X., Tian, M., Kong, S., Wu, L. and Yu, J. (2020) A modified YOLOv3 detection
method for vision-based water surface garbage capture robot, International Journal
of Advanced Robotic Systems, SAGE Publications Inc., 17(3).

10. Liang, Y., Liu, C., He, T., Wang, L., Luo, Y. and Hou, M. (2022) Design of water
surface collection robot based on deep sea cage culture, In Journal of Physics:
Conference Series, IOP Publishing Ltd.
12 MSA ENGINEERING JOURNAL
Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)
 11. Naicker, H. S., Srivastava, Y., Pramod, A., Ganatra, N. P., Sood, D., Singh, S. and
Guruviah, V. (n.d.) Water Care: Water Surface Cleaning Bot and Water Body
Surveillance System,.

12. Phirke, S., Patel, A. and Jani, J. (2021) Design of an autonomous water cleaning bot,
In Materials Today: Proceedings, Elsevier Ltd, pp. 8742–8747.

13. Raghavi, R., Varshini, K. and Kemba Devi, L. (2019) International Journal of
Advanced Research in Electrical, Electronics and Instrumentation Engineering Water
Surface Cleaning Robot, [online] Available at: www.ijareeie.com.

14. Rahmawati, E., Sucahyo, I., Asnawi, A., Faris, M., Taqwim, M. A. and Mahendra,
D. (2019) A Water Surface Cleaning Robot, In Journal of Physics: Conference Series,
Institute of Physics Publishing.

15. Wang, P., Fu, Y., Yang, H., Peng, H., Wang, H. and Ding, R. (2022) Structure design
and simulation analysis of the water surface cleaning robot, In Journal of Physics:
Conference Series, IOP Publishing Ltd.

16. Yang, H. and Kim, C. (2021) Research related to the Appearance Design of Water
Surface Cleaning Robot based on Bionics, In IOP Conference Series: Earth and
Environmental Science, IOP Publishing Ltd.

13 MSA ENGINEERING JOURNAL

Volume 2 Issue 2, E-ISSN 2812-4928, P-ISSN 28125339 (https://msaeng.journals.ekb.eg//)