An autonomous rover for efficient waste segregation and collection in urban parks and fields

Abstract

This project presents the design and implementation of an autonomous rover system for efficient waste segregation and collection in urban parks and fields. Bangladesh faces significant challenges in solid waste management due to rapid urbanization and limited infrastructure which results in environmental and health hazards. To address this, we developed a rover equipped with a Raspberry Pi 5, Pi camera and a 4-DOF robotic arm. A YOLOv8n machine learning model enables real-time waste detection and classification into biodegradable and non-biodegradable categories with over 84% accuracy from practical scenarios. The rover follows a zigzag navigation path where it autonomously detects and picks up the waste which is then deposited into a compartmentalized bin that rotates through a servo. Extensive simulations and tests were conducted which validated its efficiency and achieved complete field coverage within 20 minutes while maintaining sustainable power usage. This solution demonstrates a scalable, cost-effective, and eco-friendly approach to urban waste management which promotes cleaner environments and improved public health.