Development of a Microcontroller-Based Waste Sorting System for Metal and Non-Metal Classification Using an Inductive Proximity Sensor

Abstract

The increasing volume of waste requires efficient sorting approaches to support recycling and sustainable waste management. Manual waste separation is labor-intensive, time-consuming, and susceptible to classification errors. This study develops a microcontroller-based waste sorting system for distinguishing metal and non-metal waste using an inductive proximity sensor as the primary detection device. Two HC-SR04 ultrasonic sensors are integrated to monitor the fill levels of metal and non-metal waste bins, while a servo motor directs the classified waste to the corresponding containers. An LCD module provides real-time information on waste classification and bin capacity. The system was developed using a Research and Development (R&D) approach and evaluated through functional testing, waste classification accuracy testing, and bin capacity monitoring. The experimental results showed that the system achieved a metal waste classification accuracy of 53.8% and a non-metal waste classification accuracy of 100%, resulting in an overall classification accuracy of 73.9%. The results indicate that the inductive proximity sensor provides reliable detection of non-metal waste classification but has limitations in detecting metal-containing objects, particularly those affected by surface covering, material characteristics, and object positioning within the sensor's effective detection range. These findings demonstrate the feasibility of the proposed prototype for small-scale and educational waste sorting applications while highlighting the need for further improvement in sensor placement and mechanical feeding design to improve metal detection reliability.