Optimization of Waste Management Disposal Using Edge Impulse Studio on Tiny-Machine Learning (Tiny-ML)
Chinomso Daniel Okoronkwo, Charles Ikerionwu, Visham Ramsurrun, Amar Kumar Seeam, Nkechi Faustina Esomonu, Victory Obodoagwu · 2024
This study explores the implementation of Tiny Machine Learning (TinyML) for object detection to automate and trigger events based on inferences. The system integrates a camera module with an Arduino Nano BLE Sense microcontroller. The system leverages Edge Impulse for the machine learning model. The primary objective is to enhance waste management efficiency by ensuring that refuse bins open only when the correct type of waste is presented. This automation aims to reduce contamination in recycling streams and improve the accuracy of waste sorting. This research includes the design and training of the TinyML model, integration with the camera, microcontroller and real-world testing to evaluate the system's accuracy and responsiveness. Preliminary results indicated 93.38% accuracy in waste object detection and significant improvements in waste management processes. This approach demonstrated the potential of TinyML in creating smart, sustainable solutions for environmental challenges.