A Novel Method for Waste Prediction Using Multilayer Convolutional Neural Network with IoT Based Intelligent Waste Management System

To eliminate the need for constant human monitoring and improve trash management in large cities, a prototype system has been created. In this case, the issue of disorganized and unsystematic garbage collection is addressed by using an embedded IoT system to track the quantity of garbage placed in each container. Wet and dry garbage, as well as electronic waste, may be separated with the use of an automated system (a mechanical and software arrangement). Wet and dry garbage are separated using sensors. An infrared (IR) sensor can detect the presence of both wet and dry trash. Next, a moisture sensor is used for recognizing damp trash, and two machine learning models, a multilayer perceptron, and a multilayer convolutional neural network (ML-CNN), are utilized for sensing electronic waste. Binary categorization, such as dry or wet and electronic trash, is provided by the multilayer perceptron, and the ML-CNN detects the different waste classes. Furthermore, a camera is positioned in front of the garbage conveyor belt to photograph the trash for sorting purposes. After being sorted, trash is deposited using an electronic controller into the properly designated container. At the very top of each container is a set of ultrasonic sensors. The trash can's capacity is determined by the readings from the ultrasonic sensor. This allows for accurate tracking of how much trash is being stored in each container. A notification will be sent to the designated recipient once the container is full. The trash can is very adaptable.