AQUABOT: AI-Powered Robotic Waste Collector

Abstract

River pollution, caused by floating plastic and solid waste, poses a serious threat to both aquatic ecosystems and human health. Manual waste collection is inefficient, unsafe, and time- consuming. This paper presents a machine learning-enabled robotic river cleaning system designed to detect, collect, and monitor floating waste using a combination of Raspberry Pi, ESP32, camera-based object detection, and robotic arm control. The system operates using a DC adapter-based power supply and supports manual and autonomous operation modes. A camera mounted on the robot provides live video streaming and enables object detection using a trained deep learning model. Detected waste objects are collected using a servo-controlled robotic arm. The system can be remotely controlled through a mobile application interface providing movement controls, live video feedback, and robot status monitoring. This approach minimizes human involvement, enhances efficiency, and offers a scalable solution for intelligent river cleaning.