Machine Learning-Enabled Robotic Trash Collector

Abstract

India, being a holy country, experiences significant water pollution during festivals such as Ganpati-Visarjan, Navratri, and Durga Puja due to the disposal of waste into nearby water bodies. This pollution poses a serious threat to aquatic life. To address this challenge, this research presents the design and construction of a Machine Learning-enabled Robotic Trash Collector (MLRTC) that can remove floating trash from river banks, lakes, or ponds. The MLRTC is specifically designed to collect surface-level garbage, particularly non-biodegradable waste such as plastics that predominantly float on water bodies. The system integrates an onboard webcam, a Bluetooth sensor module for controlling, YOLOv3 algorithm for object detection, a motor driver for enhanced power output and precise control, and sunboard foam sheet for the prototype's construction. Some major challenges that we have faced while working on this robot are: first, the robot's reliance on solar power that is it may face difficulties in rainy or low light conditions. Over time, batteries may degrade and lose capacity, reducing the robot's lifespan and efficiency. Lakes can have unpredictable water conditions like waves or high currents which can affect the robot's stability. Also, there are obstacles like rocks and aquatic animals and therefore the robot needs to have reliable sensors to detect these obstacles in real-time. Even with automation, some human intervention is maintenance and empty debris collection units. The inclusion of the webcam and YOLOv3 algorithm enhances the robot's capability to detect and navigate around obstacles efficiently. The paper outlines the design, construction, and preliminary testing of the MLR TC, showcasing its potential for broader applications beyond simple pool cleaning. The integration of these technologies underscores the effectiveness of AI -driven robotic systems in combating environmental pollution, thereby contributing to sustainable waste management practices.