Design and implementation of a user-level, real-time smart waste management system for university

Abstract

Waste management is one of the most serious issues of contemporary campuses and urban areas that may lead to an unsanitary environment, resource wastage and environmental damage. This project introduces the design and implementation of an intelligent waste management system, which combines the use of IoT sensors, machine learning, and cloud-based monitoring. Two design dimensions were taken into account, a low-cost sensor-based design and a machine learning design with a Raspberry Pi and YOLOv8n classifier. The latter was found to be better by comparison, with 92% classification accuracy, consistent hazard detection, and real-time monitoring. Actuator control and power consumption were verified through MATLAB and Proteus simulations, and a physical prototype tested proved to be effective in segregation and optimal collection. The system minimizes the number of trips to collection, promotes recycling and allows responsible user behavior by a reward system. In keeping with the global sustainability objectives, the project will provide a replicable and scalable model of cleaner, smarter, and more sustainable communities.