The Internet of Things (IoT) encompasses a system of interconnected devices, with millions of things being connected every month. IoT’s global market, spanning several sectors like health, banking, and education, is projected to reach $1.1 trillion by 2023. One area where IoT is making an impact is waste management.

This paper proposes a smart waste management system that combines IoT and deep learning for optimal garbage management. The system intelligently classifies bio waste and non-bio waste through image classification using deep learning. It also includes an architectural development process for a smart trash box, equipped with an ultrasonic sensor, load measurement sensor, and micro-controller. This enables real-time waste monitoring through both Bluetooth communication and IoT technology via an Android application.

The proposed methodology consists of two parts: waste classification using a convolutional neural network (CNN) and the architectural design of smart trash boxes for real-time data monitoring. The classification of waste into proper categories is crucial for identifying reusable materials, allowing their utilization without degradation. The waste classification technology helps in attaining waste categories from images, and the trash box architecture uses multiple sensors for data transmission and monitoring.

The research showcases a real-time waste monitoring system utilizing deep learning and IoT. The waste classification accuracy achieved is 95.3125%, and the System Usability Scale score for user satisfaction is 86%. However, there are limitations, such as working with only five categories of indigestible waste, using only two sensors, and detecting holes in the trash box when it’s not completely full. Future improvements will address these limitations, aiming for more optimal results in waste management.

‍

References

• Research Paper (King Saud University)