A Low-Cost Autonomous Surface Vehicle (ASV) for Plastic Waste Collection on Water Bodies

Indiscriminate waste disposal worldwide is a pervasive problem for different governments. Most of this plastic waste ends up in drainage and other water bodies, posing a threat to humans and marine life. Poor plastic waste disposal clogs drainages, contributing to perennial flooding experienced in some countries. The financial implications of solving the plastic waste problem have made it difficult for the government to find a permanent solution. Thus, this project proposes a more affordable way of collecting plastic waste on water bodies with an Autonomous Surface Vehicle (ASV) that can be operated remotely through a website or autonomously. The ASV’s hulls were built from wood and reinforced with fibreglass using Computer Numerical Control (CNC) machines. A conveyor belt for collecting plastic waste was made from aluminium profiles. A Raspberry Pi and wireless network were used to create a Local Area Network (LAN), allowing for remote control of the ASV. A classification model was deployed that enabled object detection and tracking of plastic waste. Testing the ASV’s ability to identify plastic waste in a towing tank, a swimming pool, and a lagoon showed that the machine learning model had an accuracy score of 80%. The conveyor belt mechanism also picked up different types of plastic waste in all test scenarios. The project was mostly successful, and its objectives were met.