2024 International Conference on Intelligent Algorithms for Computational Intelligence Systems (IACIS)

Design and Development of an Intelligent Wireless

Pond/Lake Cleaning Robot
2024 International Conference on Intelligent Algorithms for Computational Intelligence Systems (IACIS) | 979-8-3503-6066-0/24/$31.00 ©2024 IEEE | DOI: 10.1109/IACIS61494.2024.10721954

1st Premkumar M 2nd Basavaraja Desai 3rd Bhoomika Mylar

Department of Electrical and Department of Electrical and Department of Electrical and
Electronics Engineering Electronics Engineering Electronics Engineering
Dayananda Sagar College of Dayananda Sagar College of Dayananda Sagar College of
Engineering Engineering Engineering
Bangalore, India Bengaluru, India Bengaluru, India
[email protected] [email protected] [email protected]

4th Chaithanya B K 5th Monish G

Department of Electrical and Electronics Engineering Department of Electrical and Electronics Engineering
Dayananda Sagar College of Engineering Dayananda Sagar College of Engineering
Bengaluru, India Bengaluru, India
[email protected] [email protected]

Abstract—This River pollution is a major environmental lakes, and other water bodies, ultimately contributing to a
problem that has serious problems for human health, aquatic reduction in water pollution. Rivers in India are deeply
life, and the overall health of ecosystems. Floating garbage is a important culturally and religiously, serving as essential
complicated problem to address, as it can be difficult to collect sources of drinking water, irrigation, bathing, and sites for
and remove from rivers and other waterways. Wireless river- religious rituals. The drive to clean and safeguard these rivers
cleaning robots can potentially be a valuable tool for addressing stems from a sense of duty to protect sacred natural resources
this problem. These robots can be equipped with a variety of and uphold cultural and religious values. This system
sensors and actuators to detect and collect floating garbage, significantly reduces the challenges associated with manual
without the need for human efforts. This can help to reduce the
clean-up efforts. By removing debris, the system contributes
cost and environmental impact of river cleaning operations,
while also improving the effectiveness of these efforts. This
to a reduction in water pollution. Floating plastic waste,
paper aims to design and develop a wireless river-cleaning robot discarded packaging materials, and other pollutants are
that is capable of autonomously detecting and collecting floating prevented from sinking or spreading further. The system also
garbage. The robot will be designed with a variety of sensors, plays a crucial role in preventing harm to aquatic animals. The
including cameras, ultrasonic sensors, and water quality robot floats on the water's surface. Lakes and ponds are
sensors. Image processing will be used to identify floating turning into landfills. Initiatives like this aim to restore and
garbage, as well as to avoid obstacles and navigate the river protect these valuable ecosystems. Various problems occurred
safely. The robot will also be equipped with a collection system on the surface of the pond, and river due to the dumping of
that can collect floating garbage and other debris. The robot will waste into the water directly. Here are some of the problem
be controlled wirelessly using image processing and data will be statements. Water pollution disturbs ecosystems and reduces
stored on the website. This paper helps to reduce the water biodiversity by endangering aquatic life, such as fish, plants,
pollution on floating bodies and saves the lives of aquatic and other organisms. Exposure to toxins or ingestion of
animals. contaminated water can put humans at risk for illnesses and
other negative health outcomes. A lot of contaminated water
Keywords—wireless river cleaning robot, image processing, bodies are situated in inaccessible or remote areas, making it
water quality sensor, garbage, Arduino mega, raspberry pi. difficult for manual cleaning teams to get to and adequately
I. INTRODUCTION clean these areas. The inability of current cleaning techniques
to monitor in real-time makes it more difficult to identify and
This A wireless river-cleaning robot is a device that can be react quickly to pollution incidents. Therefore, to address this
used to clean rivers, and ponds, without the need for human problem, we need the help of robots to clean the river, which
intervention. The wireless river cleaning robot aims to address should have various features to collect the debris and move it
the issue of waste debris in water bodies. Rivers in India are on the surface of the water. Here is the robot with such a
deeply important culturally and religiously, serving as feature of image processing to identify the debris and collect
essential sources of drinking water, irrigation, bathing, and the debris using the convey belt. This robot works both
sites for religious rituals. The drive to clean and safeguard manually and automatically. This robot is user-friendly to the
these rivers stems from a sense of duty to protect sacred customers. The article is structured in this context as follows:
natural resources and uphold cultural and religious values. It Section II describes about the literature review. The proposed
consists of a waterwheel-driven conveyor mechanism methodology analyzed in Section III. Technical details of the
designed to collect and remove the garbage, plastic waste, and paper are discussed in Section IV. The robot testings are done
other debris from water surfaces. By doing so, it reduces the in Section V. Mathematical calculations are discussed in
challenges associated with manual debris collection. The Section VI. The results and analysis of different operating
Indian government has made commitments to clean up major conditions are given in Section VII. Section VIII concludes
rivers such as the Ganga and Yamuna under initiatives like the the paper. Lastly, Future scope is presented in Section IX.
Namami Gange program. Deploying river-cleaning robots
aligns with these commitments and demonstrates a proactive II. LITERATURE REVIEW
approach to addressing water pollution and environmental The author [1] planned the robot which is a worked Stream
challenges. The system lifts surface debris from rivers, ponds, Cleaning Machine with the trust that it would be exceptionally

979-8-3503-6066-0/24/$31.00 ©2024 IEEE

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on February 16,2025 at 13:43:13 UTC from IEEE Xplore. Restrictions apply.
 much prudent and accommodating for stream and Lake analyze the live feed to identify waste and debris, allowing for
cleaning. The author [2] with the implementation of Wi-Fi targeted cleaning. Through iterative development and testing,
communication with the robot for collecting the information the system is optimized for efficiency and effectiveness in
about the garbage collected. The author of [3] Swachh pollution control.
believes the robot will be a potential alternative to surface
water trash collection in semi-autonomous robots in the This robot aims to autonomously identify and collect
existing system the articles use the manual control of river floating garbage without human involvement. The robot can
cleaning robots. The author of [4] research works on navigate the river, detect obstacles, and efficiently collect
deploying a solution to the unfavorable effects of floating floating debris. This paper incorporates various sensors, such
refuse and henceforth reduce the pollution cost due to as cameras, ultrasonic sensors, water quality sensors, Arduino
unsettled waste. Mega, Raspberry Pi 4, and image processing technology. The
Arduino mega controls the robot's motors to move the robot
The author of [5] It involves developing a reinforcement towards the debris. The Arduino Mega also controls the
learning framework using Q-Network to solve the given path robot's cleaning unit to collect the debris. The Humidity
planning problem of the previous numerical and graphical Sensor is used to measure the humidity level of the air. The
simulations performed and compared with various sampling Turbidity Sensor is used to measure the turbidity of a liquid.
based path planning algorithms. The author of [6] Cleaning The Air Pressure sensor is used to measure the air pressure.
waste by utilizing manual procedure would cause insufficient The GPS module is used to track the location of the robot. The
cleaning of water so a new generation of river cleaning robots. obtained sensor values are updated on to the cloud. The data
The author of [7] The robot is designed for cleaning the water will be sent to the website then the user can obtain the required
debris floating on the water bodies by using the RF technology information about the robot. In addition, for the detection of
we can collect the different types of floating waste up to 5 kgs. objects, we implemented an ultrasonic sensor to detect the
The author of [8] used IOT technology for river cleaning objects in its path and if any objects are present then this
robots. They used an IOT-based monitoring system which sensor sends signal to the Arduino Mega then the signal goes
provides connectivity with our system worldwide and GPS to the Raspberry Pi 4 this starts the image processing
will provide better navigation in the river. The author of [9] mechanism through Pi camera. The proposed system consists
Used image segmentation to predict the object here tensor of a camera that will give live surveillance, so live surveillance
flow library is used to predict whether the object is is a significant breakthrough in the field of river cleaning. The
biodegradable waste or non-biodegradable waste. use of open CV (SSD algorithm) in Raspberry Pi 4 for
accurate and efficient object detection if the detected object
The author of [10] proposed a water body cleaning robot contains waste, then the waste will be collected by the
that is controlled through a website along with the IOT cleaning unit through the conveyor belt or else if the detected
technology. The author of utilized the equipment stage with object contains obstacles like trees or stones then the
Raspberry Pi and ESP32 as the center module and the other is Raspberry Pi 4 sends a signal to the Arduino mega to change
the computer program stage utilizing CNN calculation. The the robot's direction.
author of utilized the equipment stage with Raspberry Pi and
ESP32 as the center module and the other is the computer IV. TECHNICAL SETUP
program stage utilizing CNN calculation. The author of
utilized the equipment stage with Raspberry Pi and ESP32 as The wireless river cleaning robot includes Arduino Mega
the center module and the other is the computer program stage receives the sensor’s data and communicates with Raspberry
utilizing CNN calculation. Pi. The camera and AI image processing algorithms are
integrated with Raspberry Pi. The ultrasonic sensor detects
III. METHODOLOGY objects in its path. The GPS module provides precise location
information about the robot. The L298n motor driver drives
The proposed system involves integrating live the DC motor to help propel the robot. The camera could be
surveillance and remote-control capabilities to effectively represented as live surveillance of the river-cleaning robot.
address pollution in water bodies. The system includes an The water quality sensors are Humidity, Temperature
Arduino controlled device equipped with motors for Turbidity, and Air pressure sensors to monitor the data on the
movement and a conveyor belt for waste collection, as well as water surface. Arduino mega, Sensors and Raspberry Pi
an onboard camera for live footage. Commands sent from a typically operate at 5V, the buck converter steps it down to 5V
mobile device via Bluetooth communication instruct the to prevent damage to the device. HC-05 Bluetooth module.
robot's actions, such as moving forward, turning, or starting acts as a bridge to control the robotic arm movement and
the conveyor belt operation. Object detection algorithms wheels and conveyor belt operations.

Fig. 1. Block diagram

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on February 16,2025 at 13:43:13 UTC from IEEE Xplore. Restrictions apply.
 Fig. 3. Weight withstands the capacity of the robot

Sensor Calibration: The water quality sensors are

interfaced with Arduino Mega, and the sensor data is
calibrated to ensure accurate readings. Water quality sensors
such as Humidity, Turbidity, Air Pressure, and Temperature
Fig. 2. Technical setup are calibrated and displayed on a serial monitor.
Live Surveillance Test: For live surveillance footage we
V. TESTING are using the Rasp Controller application. That live
The robot tested on various parameters such as: surveillance will be very clear and effective. Autonomous
Navigation Tests: The robot is placed in a controlled water
Weight withstanding capacity test: The robot can environment to test its autonomous navigation. Its ability to
withstand up to 11kg of weight (including the components), avoid obstacles using the ultrasonic sensor and adjust its path
whereas 7kg of waste can be collected by the robot, and without human intervention is evaluated. Bluetooth Interface:
remains weight is for the body and components in the robot.

Fig. 4. Sensors calibration detail

The Bluetooth module HC05 interfaced with Arduino of sensors, mechanical arms, and AI algorithms to navigate
Mega, to control the robot's directions, conveyor belt through waterways, and also receive water quality Table. 1.
operations, and shifting the robot from manual to automatic The components and their power consumption monitoring
mode by giving the commands from mobile. In automatic data. To implement a live surveillance system that allows us
mode, the robot will be responsible for detecting objects by to view real-time footage from the robot's onboard camera in
image processing and collects in its bin. our mobile device. By that live surveillance, we will control
manually and clean up the water surface by the conveyor belt.
The Bluetooth remote control was tested, including the
responsiveness to the various commands and the quality of the
live surveillance system. Figure 3 shows the Bluetooth and
sensors interface with Arduino Mega. Cleaning Mechanism
Test: The conveyor belt operation is tested, as blade edges
collect the floating garbage or debris.
The river cleaning robot is a specialized robot designed to
clean up rivers and other water bodies from pollutants, debris,
and waste materials. These robots typically use a combination
Fig. 5. Navigation testing on water

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on February 16,2025 at 13:43:13 UTC from IEEE Xplore. Restrictions apply.
 Table 1 shows the power consumption of each component states that force equals mass times acceleration. Here, the
in our system. The maximum current is drawn by DC motors, acceleration due to gravity acts as the acceleration.
relay, and motor driver. Because they play a major role in this
system. According to that power consumption, we selected a
particular battery as a maximum current.

TABLE I. SHOWS THE POWER CONSUMPTION OF COMPONENTS

Sl. Components Consumption Power
No. Ratings Ratings
01 Arduino Mega 5V, 300mA 1.5W
02 4 DC Motors 5V, 400mA 2.0W
03 L298n Motor Driver 9V, 300mA 1.8W
04 Turbidity Sensor 5V, 20mA 0.1W
05 DHT 11 5V, 20mA 0.1W
06 Air Pressure Sensor 5V, 15mA 0.075W
07 Ultrasonic Sensor 5V, 10mA 0.05W
08 Relay 5V, 140mA 0.7W
09 Bluetooth 5V, 30mA 0.15W
Total Current 1235mA
10 Battery 9V, 3.8Ah 34.2Wh
Fig. 6. Graphical representation of power vs time
A. Inefficiencies in Power Conversion
VI. RESULT AND ANALYSIS
During power conversion (charging and discharging),
energy is lost as heat due to inefficiencies in the conversion Movement and Navigation: The robot's ability to move
process. These losses are usually accounted for by considering forward, and backward, and turn it accordingly. The remote
an efficiency factor, which is typically between 80% to 90% commands (1, 2, 3, 4, 9) control the robot's movements such
for most battery systems. as forward, Right, Left, Stop, and Backward this manual
control operation by remote control. In automatic modes
1) Actual battery Capacity (commands 5 and 6) are evaluated. The precision of
Nominal Voltage: Typically, 3.7V for a cell. movement adjustments, based on object detection and
Lithium-ion Battery capacity=3800mAh positioning is also assessed.

=( )∗
( ) (1)

= 3800mAh*0.85

= 3230mAh

2) Theoretical Battery Life

!" #$% & ' ( %

ℎ = (2) Fig. 7. Robotic movement in the water surface
)* ! & $$#+ &*+, -#.

32301 ℎ SSD algorithm: When comparing the SSD (Single Shot

ℎ = Multibox Detector) algorithm with another object detection
12351 algorithm, one commonly used alternative is YOLO (You
Only Look Once). SSD divides the image into a grid and
ℎ ≅ 2.616 Hours predicts bounding boxes and class probabilities for these
boxes.
3) Buoyancy

5 6 7 8 9ℎ ( 9) =
: ;; ℎ 7 ∗ < = (3)

9 = 11.07 ∗ 9.81

9 = 108.59B

This equation calculates the force exerted by the weight of

an object. The force ( 9) is equal to the mass of the object
multiplied by the acceleration due to gravity (9.81 m/s² on
Earth). So, if the mass of the robot is 11.07 kilograms, then the Fig. 8. Graphical Representation of distance vs time
force exerted by its weight is 108.59 Newtons (B) . This
equation comes from Newton's second law of motion, which Obstacle Avoidance: The performance of the ultrasonic
sensor in detecting obstacles and the subsequent actions taken
by the robot to avoid them are examined.

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on February 16,2025 at 13:43:13 UTC from IEEE Xplore. Restrictions apply.
 Image Processing Efficiency: The effectiveness of the achieved as the controlling of the system is through a
SSD algorithm in real-time object detection, classification, Bluetooth controller. One of the paper's key features is the live
and position determination is analyzed. The accuracy of surveillance system, which provides real-time footage of the
detecting specific objects, such as water bottles, and the robot's surroundings. This enables users to remotely control
system's ability to distinguish between different types of the robot's movements based on the observations made
debris is crucial. Image processing will be happening in auto through the surveillance footage, allowing for targeted
mode. pollution mitigation efforts. This ensures effective waste
removal and contributes to the overall cleanliness of the water
body.
VIII. FUTURE SCOPE
Future work involves developing better encoding methods
to handle a variable number of obstacles. The collection of the
object is split into individual objects by using a separator. To
increase the battery life and range of wireless river cleaning
robots, it is imperative to solve issues related to energy
efficiency and battery power. Advances in artificial
intelligence and machine learning could enable robots to
Fig. 9. Image Processing Using Raspberry Pi prioritize areas with higher pollution levels, optimize cleaning
routes, and react in real time to changing environmental
SSD may struggle with very small objects compared to conditions.
YOLO.YOLO is renowned for its speed, often outperforming
SSD in terms of raw processing speed because it looks at the REFERENCES
entire picture in one evaluation. This can make YOLO faster [1] G Aswini. R, Rihana Begum. A. R, "An Improvised Smart Bin
in practice. YOLO v3 and above have made significant Management System Using an Object Recognition Method," 2020
International Conference on System, Computation, Automation and
improvements in detecting small objects, often surpassing Networking (ICSCAN), Pondicherry, India, 2020, pp. 1-6, doi:
SSD. SSD's accuracy is generally high, and it tends to perform 10.1109/ICSCAN49426.2020.9262333.
well across a range of object sizes. [2] P. Kanade, P. Alva, J.P. Prasad, and S. Kanade, "Smart Garbage
Monitoring System using Internet of Things (IoT)",2021 5th
Water Quality Monitoring: It involves the physical, International Conference on Computing Methodologies and
chemical, and biological characteristics of water. The Communication (ICCMC), Erode,India, 2021,pp. 330-
accuracy and reliability of the water quality sensors are 335,doi: 10.1109/ICCMC51019.2021.9418359.
analyzed by comparing the sensor data. It typically includes [3] Anas, M. M., Athiram, M., Suresh, A., Archana, K., & Shaji, M.
regular testing of parameters like turbidity, humidity, and (2022). “Water Cleaning Bot with Waste Segregation Using Image
temperature air pressure on the water surface. This data will Processing”. In Lecture notes in electrical engineering (pp. 161–171).
doi: 10.1007/978-981-19-1742-4_13.
be displayed on the serial monitor.
[4] Z. Mamatnabiyev, "Animal Tracking System Based on GPS Sensor
YOLO applies a single neural network to the full image, and LPWAN,” International Conference on Smart Information
which then divides the image into regions and predicts Systems and Technologies (SIST), Nur- Sultan, Kazakhstan, 2022, pp.
1-4, doi: 10.1109/SIST54437.2022.9945724.
bounding boxes and probabilities for each region. The
[5] Mahadev Kumbar, Aishwarya Betageri, Chandan Nerli, Muskan
predictions are then consolidated to produce the final Shaikh, 2022, “Waste Collection from Stagnant Water Bodies and
detection results. SSD is known for its speed and efficiency, Segregation using Artificial Intelligence and Raspberry Pi”,
particularly when detecting objects across various scales due International Journal of Engineering Research and Technology
to its multi-scale feature maps. It can process images quickly, (IJERT) Volume 11, Issue 11 (November
2022), doi:10.17577/IJERTV11IS110118.
which is crucial for real-time applications. Cleaning
Effectiveness: The efficiency of the conveyor belt system in [6] J. Yerramsetti, D. S. Paritala and R. Jayaraman, "Design and
Implementation of Automatic Robot for Floating Solar Panel Cleaning
collecting different sizes and types of debris is measured. The System using AI Technique," 2021 International Conference on
capacity of the collection unit and the time taken to fill it are Computer Communication and Informatics (ICCCI), Coimbatore,
considered. India, 2021, pp. 1-4, doi: 10.1109/ICCCI50826.2021.9402482.
[7] Simi P Thomas, Aswathi T, 2023, “Solar Powered Water Body
Cleaning Robot”, International Journal of Engineering Research and
Technology (IJERT) Volume 11, Issue 01 (June 2023), doi:
10.17577/ICCIDT2K23- 130.
[8] Bharath B P, Srivani E N, Bharath S, Akhil V Narayan, “GPS Guided
River Cleaning Robot”, International Journal of Research Publication
and Reviews, Vol 4, no 5, pp 4181- 4185 May 2023, doi: ISSN 2582-
7421.
[9] M. N. Mohammed, S. Al-Zubaidi, S. H. Kamarul Bahrain,M.
Zaenudin and M. I. Abdullah, "Design and Development of River
Cleaning Robot Using IoT Technology," 2020 16th IEEE International
Fig. 10. Collecting the floating garbage Colloquium on Signal Processing & Its Applications (CSPA),
Langkawi, Malaysia, 2020, pp. 84-87, doi:
VII. CONCLUSION 10.1109/CSPA48992.2020.9068718.
[10] Satheesh, J, Nair, A. P, Devipriya, M, Chithra, A, Mahesh, G, &
A wireless river-cleaning robot has been presented in this Jayasree, P. R. (2020). “Wireless Communication based Water Surface
paper. In the proposed techniques, objects are detected and Cleaning Boat”,2020 4th International Conference on Trends in
predicted with the help of the camera. The image processing Electronics and Informatics (ICOEI) (48184).
is done to predict the object. The waste collected can be doi:10.1109/icoei48184.2020.9142960.

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on February 16,2025 at 13:43:13 UTC from IEEE Xplore. Restrictions apply.