A
Sea Trekker: A Dual-Mode USV for Autonomous Floating Debris Collection Using YOLO and UWB LocalizationNaveenChandrasekar1✉Emailnaveen.chandrasekar@dxb.manipal.edu
BilalBaslar
Student Member, IEEE)
1MedwinSajan
Student Member, IEEE)
1GanesanSubramanian1,3
SanjitSriprasanna
Student Member, IEEE)
1MuhammedHamid
Student Member, IEEE)
1MohamedHanif
Student Member, IEEE)
1ShaikhMohammedAyan
Student Member, IEEE)
1AbhinavArunkumar
Student Member, IEEE)
1SyedaMaizah
Student Member, IEEE)
1OwaisAsifShaikh
Student Member, IEEE)
1GaganPramod
Student Member, IEEE)
21Department of MechatronicsManipal Academy of Higher EducationDubaiUAE
2Department of Mechanical EngineeringManipal Academy of Higher EducationDubaiUAE
3Department of Computer StudiesSymbiosis International UniversityDubaiUAE
Naveen Chandrasekar1, Bilal Baslar1 (Student Member, IEEE), Medwin Sajan1 (Student Member, IEEE), Ganesan Subramanian1, 3, Sanjit Sriprasanna1 (Student Member, IEEE), Muhammed Hamid1 (Student Member, IEEE), Mohamed Hanif1 (Student Member, IEEE), Shaikh Mohammed Ayan1 (Student Member, IEEE) Abhinav Arunkumar 1 (Student Member, IEEE), Syeda Maizah 1 (Student Member, IEEE), Owais Asif Shaikh 1 (Student Member, IEEE), Gagan Pramod2 (Student Member, IEEE)
1Department of Mechatronics, Manipal Academy of Higher Education, Dubai, UAE
2Department of Mechanical Engineering, Manipal Academy of Higher Education, Dubai, UAE
3Department of Computer Studies, Symbiosis International University, Dubai, UAE
This work was Funded and supported by Mubadala competition.
Corresponding author: Naveen Chandrasekar (naveen.chandrasekar@dxb.manipal.edu)
ABSTRACT Unmanned Surface Vehicles (USVs) are autonomous or remotely operated watercraft used in various marine applications, including surveillance, environmental monitoring, and debris collection. They offer efficient, low-risk alternatives to manned operations in challenging aquatic environments. SeaTrekker is a dual-mode USV developed by Team Globulocytosis 2.0 at the Manipal Academy of Higher Education, Dubai, for efficient aquatic debris collection. Designed with a curved catamaran-style dual-hull inspired by Bernoulli’s principle, it optimizes hydrodynamic flow while guiding debris into a passive net-based intake system. The USV operates in both teleoperated and autonomous modes, integrating YOLO-based object detection, Intel RealSense depth sensing, and UWB localization to identify and retrieve floating objects such as soda cans. Its modular aluminum structure and 3D-printed ABS hulls support sustainability and ease of assembly. During testing, SeaTrekker collected up to 22 cans per minute manually and 9 objects per minute autonomously, reaching speeds of up to 10 km/h. Challenges such as water leakage and localization drift were addressed with targeted improvements. SeaTrekker won First Prize in the Teleoperated Mode at the Mubadala Innovation Challenge 2025, highlighting its technical excellence. Future enhancements aim to improve autonomy, durability, and adaptability through waterproofing, GPS integration, safer battery systems, and dataset expansion—positioning it for real-world marine cleanup missions. INDEX TERMS Autonomous navigation, Water environmental cleanup, Object detection, Teleoperated Unmanned Surface Vehicles, Ultra-Wideband localization, YOLO |
I. INTRODUCTION
The explosive surge of plastic waste and floating debris in rivers, lakes, and coastal waters has escalated into a global environmental emergency. Urban rivers and canals are frequently subjected to illegal dumping, leading to the contamination of freshwater channels and blockage of sewer systems, which contributes to urban flooding [10]. Much of the discarded waste is often seen floating on the water surface. Aquatic pollution threatens biodiversity, disrupts ecosystems, and poses serious risks to public health and local economies. Millions of tons of plastic enter water bodies each year, where they endanger aquatic life and compromise the integrity of natural resources. Floating trash, if not removed promptly, can block waterways, damage aquatic infrastructure, and degrade water quality [41] impacting agriculture, fisheries, tourism, religious and cultural practices. Cleaning trash manually is often ineffective, as it typically requires extensive effort and covers large areas [9]. Addressing this urgent issue requires innovative and scalable solutions beyond traditional, labour-intensive manual clean-up methods.
A
Autonomous systems, particularly Unmanned Surface Vehicles (USVs), offer an efficient, safe, and real-time solution for floating waste collection [1]. These robotic platforms can operate continuously with minimal human intervention, enabling wide-area surveillance and active trash retrieval even in hazardous or hard-to-reach locations. Integrating artificial intelligence and navigation systems has further enhanced their performance in dynamic aquatic environments [3].This paper presents Sea Trekker, a novel dual-mode USV designed for autonomous floating debris collection. Equipped with the YOLO (You Only Look Once) object detection algorithm, Sea Trekker identifies and classifies surface trash with high accuracy. To ensure reliable navigation even in GPS-compromised environments such as urban canals or covered reservoirs, the system employs Ultra-Wideband (UWB) localization. The USV features both autonomous and manual operational modes, making it adaptable for controlled experiments as well as large-scale deployments. This combination of intelligent sensing, localization, and operational flexibility provides a practical, scalable solution to the challenge of surface water pollution. Recent advances in USV technologies have demonstrated the potential of such platforms for water quality monitoring and waste management. A low-cost catamaran-based USV designed for agriculture has successfully measured pH, electrical conductivity (EC), and total dissolved solids (TDS) in lakes and reservoirs [1]. Another floating platform developed to meet Malaysia’s environmental monitoring needs achieved real-time water quality analysis with 99% accuracy [2]. Navigation and autonomy improvements, such as Enhanced Quantum Particle Swarm Optimization (EQPSO), have enabled optimal path planning in complex aquatic environments [3], while UAV-based dam inspections highlight the growing role of autonomous systems in infrastructure monitoring [4]. Moreover, modular and open-source solutions like the AnSweR robot exemplify how compact, low-noise USVs can support research and environmental sensing using Robot operating system based systems [5]. Various debris-collection USVs have been proposed using conveyor belts, robotic arms, and Bluetooth-controlled systems to collect floating trash in rivers and ponds [6, 7]. Newer platforms also integrate computer vision, GPS, and inertial sensors for precise, automated debris detection and collection [8]. Building upon this existing research, Sea Trekker introduces an integrated, cost-effective, and dual-mode USV that merges AI-driven object detection and UWB-based navigation for reliable autonomous operation. The goal is to enhance environmental conservation efforts by targeting floating waste in lakes, urban water bodies, and areas where human access is unsafe or impractical.
A.
Application of USV
A
Abstract
Unmanned Surface Vehicles (USVs) have demonstrated significant versatility and utility across a broad spectrum of water-related domains. Their applications span from civilian to military use, driven by advancements in autonomous navigation and environmental sensing technologies. USVs are extensively employed in water environment monitoring [23], water quality assessment [24], and various maritime operations [25], including the evaluation of water quality at catchment scales [26]. In addition, they play a crucial role in autonomous navigation and patrol missions [27], Inland waterways [42] and have been deployed in environmental initiatives such as plastic waste [28] and floating debris collection [29]. From a defence perspective, USVs support naval warfare, maritime security operations [30], and broader military applications [31]. Furthermore, their ability to operate in hazardous and hard-to-reach areas makes them ideal for water-based search and rescue missions [32].
B.
Controllers used for USW
Various control strategies have been studied for guiding USVs in autonomous debris collection tasks on the water surface. Common controllers include Proportional-Integral-Derivative (PID) [33], Iterative Learning Controllers (ILC) [34][35][36], fuzzy logic controllers [37], Model Predictive Control (MPC) [38], hybrid controllers [39] and Internet of Things [42] that combine the strengths of multiple approaches. Fuzzy logic handles uncertainties well, and MPC offers optimal control by predicting future states, but both require high computational resources. Hybrid controllers, such as those combining fuzzy logic with PID or MPC with neural networks, aim to enhance performance under varying conditions, but often increase design complexity and tuning effort. ILC improves performance over repeated tasks by learning from previous control inputs, making it suitable for structured and repetitive operations. However, its effectiveness is limited in dynamic and unstructured environments like floating debris. In contrast, the PID controller is lightweight, easy to tune, and effective for real-time error correction. It determines the necessary forward motion of the USV based on the detected position of debris. Given its simplicity, robustness, and low computational demand, PID was selected as the optimal choice for this application.
C.
USV structural design
A dual-hull design was selected as the most suitable vessel configuration to demonstrate the desired features [11]. The unique twin-hull structure provides excellent splash resistance and stability when the model glides over waves [12]. A catamaran is a dual-hull vessel that offers several advantages over a monohull design. These include superior stability, the ability to navigate shallow waters, expansive deck space, higher load-carrying capacity, greater speed, and improved motion comfort [13]. A fully enclosed dual-hull design with a multi-layered sealed structure effectively prevents water leakage. This allows for the integration of sensors and controllable vessel, marking a significant advancement over existing bots [14].
D.
Vision based design algorithm
Conventional object detection models such as discriminatively trained part-based models [15], which are inefficient. This limitation has spurred the ongoing development of advanced detection models, particularly those based on deep learning. Deep learning offers a comprehensive network architecture that excels at extracting image features, making it a key area of focus in modern object detection research. Deep learning-based object detection methods can generally be divided into two categories: region proposal-based algorithms, such as Convolutional Neural Networks [19] and Sparse Region-based Convolutional Neural Networks [20], and regression-based algorithms, including YOLO [21] [40] and the Single Shot Multi Box Detector [22]. While region proposal methods rely on generating candidate object regions before classification, regression-based approaches like YOLO treat detection as a single regression task. YOLO predicts both bounding box coordinates and class probabilities directly from the entire image in a single forward pass. This streamlined architecture enables real-time detection with high efficiency, contributing to YOLO's popularity in speed-critical applications.
II. SYSTEM DESCRIPTION
A.
CATAMARAN HULL DESIGN
SeaTrekker adopts a catamaran-style layout, featuring two parallel hulls that provide exceptional balance, buoyancy, and stability. Each hull is designed with a gentle inward curvature from front to rear, inspired by Bernoulli’s Principle. This shape helps to accelerate water flow between the hulls, reducing drag while also increasing the velocity of water at the intake zone. The streamlined form contributes to directional stability and maneuverability. The hulls are 3D-printed using ABS plastic is a recyclable material and reinforced with fiberglass to prevent warping, cracking, and long-term material fatigue. Internally, transverse ribs have been added to each hull to improve rigidity without adding unnecessary weight. The placement of the thrusters at the rear causes the front of the vessel to lift slightly during motion, improving net alignment for optimal debris intake.
A
A
Fig. 1
Design Architecture
A
A
Fig. 2
Top view of the hull connections
B.
STRUCTURAL FRAMEWORK AND MODULARITY
A
A
Fig. 3
Full structure of Sea Trekker
A
A
Fig. 4
Hull design
A
Figure 5. Top Lid designC.
ELECTRONICS ASSEMBLY
A
A
Fig. 6
Electronics Box
A
Figure 7. Electronics and WiringD.
SYSTEM CIRCUIT DESIGN
SeaTrekker’s electronic architecture is distributed between dual hull compartments and a central control unit. Each hull contains an independent propulsion setup powered by 14.8V LiPo batteries, routed through Electronic Speed Controllers (ESCs) and protected by 50A fuses and dedicated kill switches for localized shutdown. The central control box houses the Jetson Orin Nano, which processes real-time sensor data and runs vision algorithms, and a Pixhawk controller for low-level navigation and motor control. Input signals are managed through an RC receiver and signal-switching circuit to alternate between autonomous and teleoperated modes. Additional components include Intel’s RealSense D457 depth camera, UWB modules for localization, and dual cooling fans (DGI 03) to maintain thermal stability. Voltage monitors and visual indicators (LEDs) provide system feedback, while all wiring follows a modular scheme for redundancy, simplified debugging, and easy expansion.
A
A
Fig. 9
Electronics Box Components
III. SOFTWARE ALGORITHM
A.
DEBRIS DETECTION
SeaTrekker utilizes a YOLOv12-based object detection algorithm that has been customized and trained on a dataset specifically collected during field test sessions. This dataset includes real-world images of soda cans and similar floating debris to ensure realistic performance. The model runs on the NVIDIA Jetson Orin Nano platform using CUDA acceleration, which enables high-speed, real-time inference. Visual input is provided by the Intel RealSense D455 RGB-D camera, capable of streaming both color and depth data at up to 90 FPS. Once processed by YOLO, debris is identified with bounding boxes and confidence scores, with a typical inference latency of less than 30 milliseconds—providing the responsiveness necessary for autonomous debris targeting and path adjustments.
A
A
Fig. 10
YOLO12N Model Output
A
A
Fig. 11
Custom YOLO Mode with 175 EPOCH and 1000 + IMG
A
A
Fig. 12
Object Detection Framework
B.
NAVIGATION AND OBSTACLE AVOIDANCE
To ensure accurate movement within its operational environment, SeaTrekker implements a multi-layered navigation architecture that combines geofencing, real-time obstacle detection, and visual object tracking. Geofencing is established using Ultra-Wideband (UWB) anchors strategically placed around the operational area, defining virtual boundaries that prevent the vehicle from straying outside designated zones. The RealSense D455 depth camera supports obstacle detection through three key methods: depth mask analysis for identifying nearby obstructions, surface normal estimation to detect irregular or sloped surfaces, and infrared edge detection to recognize sudden spatial discontinuities that may indicate solid barriers.
A
A
Fig. 13
Obstacle Detection and Avoidance Strategy
C.
LOCALISATION AND MISSION PLANNING
SeaTrekker relies on a UWB-based localization system to determine its global position within the defined space. This is achieved by triangulating signals between the USV and a series of fixed anchors placed at known locations. The bot continually updates its coordinates, feeding this data into a mission planning algorithm that prioritizes the most efficient debris collection paths. Objects that are missed during initial passes are logged and scheduled for revisit, ensuring comprehensive area coverage. The localization and planning module also incorporates dynamic updates to avoid previously detected obstacles and optimize the collection sequence.
A
A
Fig. 14
UWB Testing in 20M*10M Area
D.
AUTONOMOUS CONTROL SYSYTEM
The autonomous functions of SeaTrekker are governed by a Proportional-Integral-Derivative (PID) control loop that dynamically adjusts the thruster output based on the position of identified debris in the camera feed. This enables smooth navigation toward targets, consistent alignment during debris intake, and reactive changes in path when necessary. Data from the YOLO detection model, the RealSense depth camera, and the UWB localization system is integrated into a unified environmental map. In this map, static elements like walls or barriers are logged as permanent obstacles, while floating debris is classified as dynamic objects that are continuously updated. This allows the bot to make context-aware navigation decisions with a high level of autonomy.
A
A
Fig. 15
Path Planning Algorithm
E.
CONTROL MODES AND SAFETY MECHANISM
SeaTrekker is capable of operating in two distinct modes. In teleoperated mode, a human user navigates the USV via a live FPV feed transmitted by the DJI O3 Air Unit, allowing for responsive manual control. In autonomous mode, the system independently follows a predefined mission based on onboard perception and planning. A signal relay circuit facilitates seamless switching between manual and autonomous operation. For safety, SeaTrekker includes physical kill switches on each hull and the central control box, as well as a remote override that can instantly halt all thruster activity. A 40A general-purpose fuse safeguards the electrical system from overcurrent incidents, ensuring both equipment and user safety during operation.
F.
DEBRIS COLLECTION MECHANISM
Sea Trekker employs a passive debris collection system that leverages the vessel’s forward motion to gather floating trash. Between the two hulls is a tensioned cuboidal net suspended on structural beams using waterproof adhesive and zip-tie mounts. At the front, a custom 3D-printed funnel structure guides floating objects directly into the net. A one-way mesh gate, built using angled zip ties, allows debris to enter but prevents it from escaping. This energy-efficient system is highly effective at capturing floating materials such as soda cans or plastic waste which relying on complex moving parts. Floatation elements on the intake assembly help keep the mechanism at the waterline, ensuring consistent contact with surface debris.
A
A
Fig. 16
Debris Collection
A
A
Fig. 17
One Way Trash Collection
IV. EFFICIENCY AND PERFORMANCE
A.
DEBRIS COLLECTION RATE
SeaTrekker demonstrates impressive collection capabilities in both operational modes. During teleoperated testing, it achieved a collection rate of up to 22 soda cans per minute. In autonomous mode, the system successfully collected up to 9 floating objects per minute, showcasing its real-time detection and path planning efficiency.
B.
POWER CONSUMPTION
The system’s power draw was carefully monitored during testing. While the calculated peak consumption reached 829 watts, the actual average usage stabilized around 418 watts under typical operating conditions, ensuring energy-efficient performance.
C.
SPEED
SeaTrekker is capable of reaching a top speed of 3 meters per second, equivalent to approximately 10.8 kilometers per hour. This allows for rapid deployment and repositioning during both manual and autonomous cleanup operations.
D.
BATTERY LIFE
At 75% thrust, the propulsion system offers up to 25 minutes of continuous runtime. Meanwhile, the onboard processing unit, powered by a 25Ah power bank, remains active for approximately 2.5 hours—enabling long-duration autonomous missions.
E.
PROPULSION SYSTEM
The USV is powered by four Blue Robotics T200 thrusters, collectively providing up to 18 kilograms of thrust. This configuration enables high maneuverability, allowing SeaTrekker to make tight turns and maintain control in confined environments.
F.
STABILITY AND CONTROL
Thanks to its dual-hull catamaran design and low center of gravity, SeaTrekker maintains excellent lateral stability. The curved hull geometry reduces drag and helps preserve a consistent waterline, which is critical for effective debris intake.
G.
COLLISION RATE
In teleoperated mode, SeaTrekker registered zero collisions, confirming high responsiveness and control accuracy. During autonomous trials, only two minor collisions were recorded per test run, both of which are being analyzed for future improvements.
H.
RESPONSIVENESS
The system includes a DJI O3 Air Unit for FPV streaming, delivering a latency of just 3 milliseconds. This near-instant feedback ensures seamless control during manual operation, especially in dynamic cleanup environments.
A
A
Fig. 18
Velocity/ Flow Simulation
A
Figure 19. Coaxial Thruster PlacementV. PID CONTROL
SeaTrekker’s real-time maneuvering relies on a finely tuned Proportional-Integral-Derivative (PID) control system that adjusts thruster outputs based on live feedback from object detection and localization inputs. This closed-loop control architecture enables the vessel to align precisely with detected debris, maintain consistent heading in turbulent water conditions, and execute smooth, calculated turns.
By minimizing overshoot and oscillation, the PID algorithm ensures that SeaTrekker remains stable during collection maneuvers and responsive during path corrections—especially crucial in narrow waterways or cluttered environments. The controller continuously receives positional offsets from the YOLO-based object detection system and UWB localization grid, converting them into precise motion commands, making SeaTrekker both agile and efficient in its autonomous missions.
VI RESULT AND DISCUSSION
SeaTrekker’s performance trials confirmed its capability as a high-efficiency aquatic cleanup system. In teleoperated mode, the USV collected up to 22 soda cans per minute, while in autonomous mode, it achieved a rate of 9 floating objects per minute. Stability was consistently maintained across all test runs, with no collisions during manual operation and only one minor incident per autonomous trial, showing effective real-time navigation and object tracking.
With a maximum speed of 3 m/s (10.8 km/h) and a combined 18 kg of thrust from four T200 thrusters, the system demonstrated strong maneuverability. Power consumption averaged 418 W, with a calculated peak of 829 W. The propulsion system provided 25 minutes of runtime, while onboard processing lasted up to 2.5 hours using a 25Ah power bank.
SeaTrekker can carry up to 30 kg of debris per mission. Operating autonomously, a single unit can perform four cleanup runs per day, removing up to 120 kg of waste daily equivalent to 3.6 metric tons per month. When deployed as a fleet of 10, this scales to 35 + metric tons of monthly cleanup, targeting plastic bottles, wrappers, and algae before they degrade into microplastics.
Additionally, each electric USV offsets approximately 4.2 metric tons of CO₂ emissions annually, eliminating the need for fossil-fuel-based boats. Field mapping showed 90–95% effective coverage, with an onboard system logging missed targets for revisit. These results confirm SeaTrekker’s potential for scalable, sustainable deployment in urban waterways, marinas, and ecologically sensitive areas.
VII CONCLUSION
SeaTrekker successfully integrates smart sensing, sustainable design, and adaptive control systems into a single platform for autonomous waterway cleanup. Its flexible dual-mode operation (manual and autonomous), real-time object detection, and efficient intake system establish it as a practical and eco-friendly solution. The project has validated the effectiveness of modular construction and cross-domain teamwork in building scalable environmental robotics.
VIII. FUTURE SCOPE
Future developments will focus on enhancing SeaTrekker’s autonomy, durability, and mission efficiency. Planned upgrades include GPS-based global deployment, trash categorization using expanded datasets, battery management systems, and solar-charged smart docks. Additional modules—such as oil spill recovery, algae removal, and coral reef monitoring—will broaden the USV’s environmental impact. These improvements will position SeaTrekker as a deployable fleet solution for urban waterways, resorts, and conservation zones worldwide.
IX. ACKNOWLEDGEMENTS
We gratefully acknowledge Mubadala for their generous support and funding of our USV project. Their grant of AED 16,131 played a crucial role in enabling our team to pursue a technically ambitious and environmentally impactful solution. This backing not only empowered us to prototype, test, and refine SeaTrekker but also reflected Mubadala’s commitment to fostering student-led innovation and sustainable technology development. We are proud to have delivered a high-performance system that aligns with their vision for future-forward initiatives.
X. AWARDS
SeaTrekker has secured 1st place at the Mubadala 2025 competition, winning a prize of 30,000 AED for excelling in the teleoperated and autonomous track. Additionally, it earned 2nd place at the IEEE 2025 Senior Design Project (SDP) competition, receiving a 2,000 AED prize. These awards recognize the innovation and technical excellence demonstrated by SeaTrekker.
A
A
Fig. 20
1ˢᵗ PLACE - MUBADALA 2025 - PRIZE − 30,000 AED TELEOPERATED TRACK
A
A
Fig. 21
2ⁿᵈ PLACE IEEE - SDP - PRIZE 2,000 AED SENIOR DESIGN PROJECT
XI. Declarations
A
A
Consent to Publish
• declaration: not applicable.
Ethics and Consent to Participate declarations: not applicable.
Funding: Mubadala.
Electronic Supplementary Material
Below is the link to the electronic supplementary material
A
Author Contribution
B.B. led the team, managed finances, handled procurement, contributed to electrical and hardware integration, and secured the grant. M.V.S. developed and led the autonomous system, configured the PID control, and provided leadership support. A.A.K. designed the hull structure, optimized weight-to-strength ratio, and co-authored the project report. S.S. set up the Jetson Orin Nano, integrated ROS2 with UWB and IMU, and built the navigation system. M.H.N. designed the electronics box, solved waterproofing issues, and contributed to CAD designs and testing. S.M. supported electronics box development, contributed to CAD designs, and optimized camera placement. O.S.A. set up UWB sensors, configured DJI systems, and provided manufacturing and morale support. S.M.A. implemented YOLO object detection, managed the Roboflow workspace, and supported the presentation. M.A.H. annotated AI datasets, conducted waterproofing R&D, and engineered the passive trash intake system. G.P. fabricated the hull structure, assisted in design feedback, and implemented the trash storage system. N.C. served as the faculty mentor for the project. G.S. supported the preparation of the proposal. All authors reviewed and approved the final manuscript.
A
A
Data Availability
The datasets generated and/or analysed during the current study are available in the following link [https://universe.roboflow.com/hehee-cd5nm/mahe-dubai-2025-usv-sta9w], and are also provided as supplementary files with this article.
A
Acknowledgement
We gratefully acknowledge Mubadala for their generous support and funding of our USV project. Their grant of AED 16,131 played a crucial role in enabling our team to pursue a technically ambitious and environmentally impactful solution. This backing not only empowered us to prototype, test, and refine SeaTrekker but also reflected Mubadala’s commitment to fostering student-led innovation and sustainable technology development. We are proud to have delivered a high-performance system that aligns with their vision for future-forward initiatives.
XII. References
1.
Hermawan Nugroho CJ, Xan TJ, Yee LK, Yong. Liew Zhan Quan, and Muhammad Ilhamdi Rusydi. Development of Unmanned Surface Vehicles (USV) to Monitor the Quality of Water for Agriculture.
2.
Zahari N, Mubin P, Sankar MH, Zawawi NT, Roslin NA, Aziz. Farah Nurhikmah, and Muhammad Husni Faiz Khamarudin. Development of Floating Platform for Water Quality Monitoring.
3.
Liu F, Wang M, Zhao B, Zhang S. An Optimization-Based Path Planning Method for Unmanned Systems in Multi-obstacle Environments. State Key Laboratory of Mechanical Transmission for Advanced Equipment, Chongqing University; 2023.
4.
UAV Dam Study. Lessons Learned: Monitoring Dam Infrastructure Assets Using Unmanned Aerial Vehicles (UAVs). 2023.
5.
Sanfilippo F, Tang M, Steyaert S. The Aquatic Surface Robot (AnSweR), a Lightweight, Low Cost, Multipurpose Unmanned Research Vessel. In Intelligent Technologies and Applications. INTAP 2020, edited by Yildirim Yayilgan, I.S. Bajwa, and F. Sanfilippo, 305–316. Communications in Computer and Information Science, vol. 1382. Cham: Springer, 2021. https://doi.org/10.1007/978-3-030-71711-7_21
6.
Akib A et al. Unmanned Floating Waste Collecting Robot. TENCON 2019–2019 IEEE Region 10 Conference, Kochi, India, 2019, pp. 2645–2650. https://doi.org/10.1109/TENCON.2019.8929537
7.
SaiTeja G. and P. S. Design and Simulation of an Autonomous Surface Vehicle for Trash Collection Using ROS. 2021 IEEE 9th Region 10 Humanitarian Technology Conference (R10-HTC), Bangalore, India, 2021, pp. 1–6. https://doi.org/10.1109/R10-HTC53172.2021.9641589
8.
Subhashini K, Shreya Shree K, Abirami A, Kalaimathy RD, Deepika R. and J. Thamil Selvi. Autonomous floating debris collection system for water surface cleaning. In 2024 International Conference on Communication, Computing and Internet of Things (IC3IoT), pp. 1–6. IEEE, 2024.
9.
Ganvir KD, Pranay R, Nerkar LW, Ghate, Harshata H. Bhagat. The impact of water pollution and preliminary study on river trash collecting mechanism. Int J Tech Res Appl. 2019;7(1):85–7.
10.
Tharani M, Amin AW, Rasool F, Maaz M, Taj M, Muhammad A. Trash Detection on Water Channels. In: Mantoro T, Lee M, Ayu MA, Wong KW, Hidayanto AN, editors. Neural Information Processing. ICONIP 2021. Lecture Notes in Computer Science. Volume 13108. Cham: Springer; 2021. https://doi.org/10.1007/978-3-030-92185-9_31.
11.
Pawara MU, Mahmuddin F. Developing a Mobile Floating Structure as an Off shore Wind Energy Harvesting System in Indonesia Sea Areas. In Proceedings of the 3rd International Conference on Advances in Mechanical Engineering (ICAME), pp. 835–841. 2017.
12.
Shao W, Shi S, Jiang T, Sang JJT, Zhao L. Research on wave test of amphibious Twin-hull aircraft model in water pond. In Journal of Physics: Conference Series, vol. 2955, no. 1, p. 012038. IOP Publishing, 2025.
13.
Srinivas DVG. and Chirapa Srinivas. Design Of Catamaran Ship Strength Deck To Withstand The Crane Load. Dasari. V. G. Srinivas,1 Chirapa Srinivas / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248–9622 www.ijera.comVol. 3, Issue 2, March -April 2013, pp.1166–1176.
14.
Shukla A, Ross R, Bhattacharya B, Stumpf A. Autonomous water sampling and quality monitoring in remote locations: A novel approach using a remote-controlled boat. HardwareX. 2025;22:e00634.
15.
Felzenszwalb PF, Girshick RB, McAllester D, Ramanan D. Object detection with discriminatively trained part-based models. IEEE Trans Pattern Anal Mach Intell. 2010;32(9):1627–45. https://doi.org/10.1109/TPAMI.2009.167.
16.
16, Uijlings J, Sande KVD, Gevers T, Smeulders A. Selective search for object recognition. Int J Comput Vis. 2013;104:154–71. https://doi.org/10.1007/s11263-013-0620-5.
A
17.Vedaldi A, Gulshan V, Varma M, Zisserman A. Multiple kernels for object detection. In: 2009 IEEE 12th International Conference on Computer Vision, pp. 606–613 (2009). https://doi.org/10.1109/ICCV.2009.5459183
18.
Gan G, Cheng J. Pedestrian detection based on HOG-LBP feature. In: 2011 Seventh International Conference on Computational Intelligence and Security, pp. 1184–1187 (2011). https://doi.org/10.1109/CIS.2011.262
19.
Girshick RB, Donahue J, Darrell T, Malik J. Rich feature hierarchies for accurate object detection and semantic segmentation. In: 2014 IEEE Conference on Computer Vision and Pattern Recognition, pp. 580–587 (2013). https://doi.org/10.1109/CVPR.2014.81
20.
Sun P et al. Sparse R-CNN: end-to-end object detection with learnable pro posals. In: 2021 IEEE/CVF Conference on Computer Vision and Pattern Recogni tion (CVPR), pp. 14449–14458 (2021). https://doi.org/10.1109/CVPR46437.2021. 01422.
21.
Yan F, Xu Y. Improved target detection algorithm based on yolo. In: 2021 4th International Conference on Robotics, Control and Automation Engineering (RCAE), pp. 21–25 (2021). https://doi.org/10.1109/RCAE53607.2021.9638930
22.
Shyam D, Kot A, Athalye C. Abandoned object detection using pixel-based f inite state machine and single shot multibox detector. In: 2018 IEEE International Conference on Multimedia and Expo (ICME), pp. 1–6 (2018). https://doi.org/10.1109/ICME.2018.8486464
23.
Zhang M, Li D, Xiong J, Yuqing He. Multi-dimensional water sampling unmanned surface vehicle development and application. Int J dynamics control. 2023;11(6):3188–208.
24.
Mendoza-Chok, Jorch JCC, Luque. Niel Franck Salas-Cueva, Daniel Yanyachi, and Pablo Raul Yanyachi. Hybrid control architecture of an unmanned surface vehicle used for water quality monitoring. IEEE Access. 2022;10:112789–98.
25.
Heng H, Ghazali MHM, Wan Rahiman. Exploring the application of ant colony optimization in path planning for Unmanned Surface Vehicles. Ocean Eng. 2024;311:118738.
26.
Katsouras G, Dimitriou E, Karavoltsos S, Samios S, Sakellari A, Mentzafou A. Nikolaos Tsalas, and Michael Scoullos. Use of unmanned surface vehicles (USVs) in water chemistry studies. Sensors. 2024;24(9):2809.
27.
Xu H, Zhang X, He J, Yang Yu, and, Cheng Y. Real-time Volumetric Perception for unmanned surface vehicles through fusion of radar and camera. IEEE Trans Instrum Meas (2024).
28.
Owusu GM, Okyere JKF, Armah SK, Peter Lawerh K. A Low-Cost Autonomous Surface Vehicle (ASV) for Plastic Waste Collection on Water Bodies. In 2024 12th International Conference on Control, Mechatronics and Automation (ICCMA), pp. 41–46. IEEE, 2024.
29.
Subhashini K, Shreya Shree K, Abirami A, Kalaimathy RD, Deepika R. and J. Thamil Selvi. Autonomous Floating Debris Collection System for Water Surface Cleaning. In 2024 International Conference on Communication, Computing and Internet of Things (IC3IoT), pp. 1–6. IEEE, 2024.
30.
Boretti A. Unmanned surface vehicles for naval warfare and maritime security. J De?F Model Simul (2024): 15485129241283056.
31.
Criollo L, Carlos Mena-Arciniega, and, Xing S. Classification, military applications, and opportunities of unmanned aerial vehicles. Aviation 28, no. 2 (2024): 115–127.
32.
Bai Y, Wang Y, Wang Z, Zheng K. Trajectory tracking and docking control strategy for unmanned surface vehicles in water-based search and rescue missions. J Mar Sci Eng. 2024;12(9):1462.
33.
Liu S, Zuo Y, Li T, Wang H, Gao X. Adaptive fixed-time PID-based control of uncertain nonlinear systems and its application to unmanned surface vehicles. Int J Syst Sci. 2024;55(14):2815–24.
34.
Tang Y, Wu F, Gong X, Chen C, Liu H, Ma J, Qin Y. Huayan Pu, and Jun Luo. An iterative learning-based integrated motion planning and control method for autonomous patrolling of unmanned surface vehicles. Meas Sci Technol. 2024;35(12):126214.
35.
Naveen C, Meenakshipriya B, Tony Thomas A, Sathiyavathi S, Sathishbabu S. Real-time implementation of iterative learning control for an electro-hydraulic servo system. IETE J Res. 2023;69(2):649–64.
36.
Naveen C, Meenakshipriya B, Sathiyavathi S, Tony Thomas A. Design and Implementation of Iterative Learning Control for an Electro-Hydraulic Servo System. J Sci Industrial Res (JSIR). 2023;82(05):579–88.
37.
Wang F, Wang C, Wang Y, Chemori A, Zhang X, Zhang K, Zhang Y. Data-driven fuzzy logic control method for improved USV path planning. J Supercomputing. 2025;81(7):1–32.
38.
Zhao B, Zhang X, Liang C, Han X. An improved model predictive control for path-following of USV based on global course constraint and event-triggered mechanism. Ieee Access. 2021;9:79725–34.
39.
Lv C, Yu H, Chi J, Xu T, Zang H. Hui lue Jiang, and Zhaowen Zhang. A hybrid coordination controller for speed and heading control of underactuated unmanned surface vehicles system. Ocean Eng. 2019;176:222–30.
40.
Li Y, Wang R, Gao D, Liu Z. A floating-waste-detection method for unmanned surface vehicle based on feature fusion and enhancement. J Mar Sci Eng. 2023;11(12):2234.
41.
Fernández Montenegro J, Manuel B, Villarini A, Angelopoulou E, Kapetanios. Jose Garcia-Rodriguez, and Vasileios Argyriou. A survey of alzheimer’s disease early diagnosis methods for cognitive assessment. Sensors. 2020;20(24):7292.
42.
Dunggat D, Nawi IM, Saeed S, Ba Hashwan MH, Md Khir I, Maisarah Sadale UI, Bature A, Izham Razali A, Ismail A, Firdaus AM, Isa, Izzani MH, Mansor A. Unmanned surface vehicle for water wastes collection. In 2023 IEEE International Conference on Sensors and Nanotechnology (SENNANO), pp. 204–207. IEEE, 2023.
A
43.Cheng Y, Jiang M, Zhu J, Liu Y. Are we ready for unmanned surface vehicles in inland waterways? The usvinland multisensor dataset and benchmark. IEEE Rob Autom Lett. 2021;6(2):3964–70.
Syeda M. a third-year Mechatronics engineering student at MAHE Dubai, excels in robotics and automation. She is skilled in AutoCAD, Fusion 360, and Shapr3D, with a talent for refining innovative ideas through hands-on problem-solving. In the award-winning Mubadala USV project, she led the design of an optimized electronics enclosure. At the Emirates Robotics Competition 2025, her color-based ball collection strategy won the Most Innovative Award. She contributed to marine robotics in the IROS and MATE ROV 2025 competitions. Currently interning at Dubai Future Foundation, she is developing a silicon-tipped gripper for transparent object handling with high precision and adaptability.
.
He works extensively in. consultancy research and plays a key role in driving innovative project development.
