Design and Development of an Autonomous, IoT-Based, Solar Powered Surface Vehicle for Algae Detection, Collection and Mitigation using Deep Learning Models

Abstract

Freshwater bodies are shrinking in both number and quality, and a large number of the remaining lakes and ponds are now being periodically impacted by Harmful Algal Blooms (HABs). These HABs cause a disruption in the natural ecosystem, make the water unsuitable for use, and require periodic manual processing. These existing techniques are all either chemical-based or involve periodic visits to a site, and this renders them inadequate for continuous monitoring. To address this need, we present an autonomous, solar-powered surface robot capable of detecting, inhibiting, and collecting algae in real-time. The platform utilizes GPS-aided navigation and an AI-based vision module that performs continuous surface scanning and updates its detection model directly on board as needed. Once an algal patch is identified, the vehicle navigates to the region, applies targeted ultrasonic excitation for non-chemical inhibition, and then activates a dedicated mechanical system to collect the resulting biomass. Initial experiments demonstrate that the system can efficiently perform these stages and with low power consumption, showcasing its promise as a practical solution for long-term lake restoration and automated algal management.

Introduction

The text addresses the growing challenge of Harmful Algal Blooms (HABs) in freshwater systems and the limitations of existing management approaches. HABs develop rapidly due to nutrient imbalance and climate change, causing oxygen depletion and toxin release that threaten ecosystems and water safety. Traditional mitigation methods—manual removal, chemical treatment, and stationary monitoring systems—are either slow, environmentally harmful, or limited in scope, typically addressing only detection or inhibition without mobility or biomass removal.

To overcome these gaps, the work proposes an integrated, solar-powered Autonomous Surface Vehicle (ASV) capable of real-time HAB detection, non-chemical inhibition, and physical algae collection within a single closed-loop system. The platform combines GPS-guided navigation, computer vision–based algae detection, ultrasonic growth suppression, and targeted biomass harvesting, enabling continuous and adaptive operation in lakes and reservoirs.

The system employs a dual-microcontroller architecture (ESP32 for vision and sensing, ESP8266 for navigation and communication), supported by cameras, GPS, water-quality sensors, solar power, and differential motor propulsion. For algae detection, the study evaluates three deep-learning object detection models—YOLOv8, YOLOv11, and Faster R-CNN—using a combined dataset of public algae images and custom images captured by the ASV. Models are assessed based on accuracy (mAP), inference speed (FPS), and model size to determine suitability for real-time, resource-constrained deployment.

Results emphasize that single-stage YOLO models offer the best balance between accuracy and speed for autonomous operation, while two-stage R-CNN models provide higher precision at the cost of slower inference. Overall, the proposed system demonstrates a novel, eco-friendly, and autonomous approach to HAB management by unifying detection, inhibition, and collection, addressing a critical gap in current freshwater monitoring and mitigation technologies.

References

[1] N. H. Ibrahim et al., “A Review on the Biological, Physical and Chemical Mitigation of Harmful Algal Bloom.” [2] D. M. Anderson, A. D. Cembella, and G. M. Hallegraeff, “Progress in understanding harmful algal blooms (HABs): Paradigm shifts and new technologies for research, monitoring and management.” [3] A. K. Misra, P. K. Tiwari, and E. Venturino, “Modeling the impact of awareness on the mitigation of algal bloom in a lake.” [4] “Prevention, Control and Mitigation of Harmful Algal Blooms: A Research Plan.” [5] D. M. Anderson, “Prevention, Control and Mitigation of Harmful Algal Blooms: Multiple Approaches to HAB Management.” [6] Arabinda Samantaray, Baijian Yang, J. Eric Dietz, and Byung-Cheol Min, ‘‘Algae Detection Using Computer Vision and Deep Learning.’’ [7] Jhony-Heriberto Giraldo-Zuluagaa, Geman Diez, Alexander G ?omez, Tatiana Mart ??nez, Mariana Pe ?nuela V ?asquez, Jes ?us Francisco Vargas Bonilla, Augusto Enrique Salazar Jim ?enez, ‘‘Automatic Identification of Scenedesmus Polymorphic Microalgae from Microscopic Images.’’ [8] NASA ARTICLE, “Applying Machine Learning to Harmful Algal Blooms.” [9] Seyit Uguz, Yavuz Selim Sahin, Pradeep Kumar, Xufei Yang and Gary Anderson, “Real-Time Algal Monitoring Using Novel Machine Learning Approaches.” [10] D. M. Anderson, “Prevention, Control and Mitigation of Harmful Algal Blooms: Multiple Approaches to HAB Management.” [11] Manhiro Flores-Iwasaki, Grobert A. Guadalupe, Miguel Pachas-Caycho, Sandy Chapa-Gonza , Roberto Carlos Mori-Zabarburú and Juan Carlos Guerrero-Abad., “Internet of Things (IoT) Sensors for Water Quality Monitoring in Aquaculture Systems: A Systematic Review and Bibliometric Analysis.” [12] Dr.R.Usha, E. Ramesh Babu, B.Uma, Dr.B.Kishori, “Development of an IoT System for Real-time Monitoring and Control of Spirulina Cultivation.” [13] Lean Karlo S. Tolentino, Sheila O. Belarmino, Justin Gio N. Chan, Oliver D. Cleofas Jr, Jethro Gringo M. Creencia, Meryll Eve L. Cruz, JC Glenn B. Geronimo, John Peter M. Ramos, Lejan Alfred C. Enriquez, Jay Fel C. Quijano, Edmon O. Fernandez,Maria Victoria C. Padilla, “IoT-based Closed Algal Cultivation System with Vision System for Cell Count through ImageJ via Raspberry Pi.’’ [14] Andy Wahyu Hermanto, Achmad Dzulfiqor, ‘‘Solar Power Generator as an Alternative Energy Source for the Propeller Driver Motor.’’ [15] Abdullah, Sikandar Ali, Ziaullah Khan, Ali Hussain, Ali Athar and Hee-Cheol Kim “Computer Vision Based Deep Learning Approach for the Detection and Classification of Algae Species Using MicroscopicImages.”

Copyright

Copyright © 2025 Shreya L, Supriya K Prasad, Swathi Bhat, Vaishnavi S Kale, Sumangala Gejje. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.