Paper Title
IoT Based Intelligent Garbage Segregation and Recycling System using Deep Learning Algorithm

Abstract
Explosive growth in the population has imposed a great hurdle in the Management, Recycling and Disposal of Garbage. People are more prone to infectious diseases. Also, the entire world is facing newer types of health hazards. Existing techniques still have a lot of limitations in management and segregation of waste in an effective manner. For alleviating the process of garbage disposal and to keep up the cleanliness it is mandatory to have a Smart Garbage Managing System. Also, newer techniques of segregation and recycling of waste in precise manner is the need of the hour. This project proposes a deep learning algorithm which is implemented with back propagation algorithm for waste segregation and management. The waste in the dustbin is detected and segregated by CNN Classification. IoT enabled environment along with Deep Learning Algorithm is used to detect and segregate the Garbage in the dustbins with the aid of Sensor devices. IR sensor is utilized to distinguish the categories of the waste material. The moisture sensor is used to analyze and report the moisture content in the waste, and if there is moisture content available then the waste is segregated into a separate Bin. Metal sensor is used to separate the metal items separately. Data collected by each smart truck is shared with the nearby Industries for the disposal of the Garbage. Deep Learning Algorithm is also used to recognize the image to separate recyclable waste. This project involves simulation of proposed Deep learning algorithm with backpropagation technique for segregation of waste such as paper box, glass and plastic. Hence the advantages of the proposed algorithm include improved computational efficiency, stable error gradient with accurate and fast learning. It is implemented and analyzed with efficient performance metrics such as accuracy, sensitivity and specificity. When simulated it was inferred that the proposed algorithm outperformed the drawbacks of the existing algorithm. Keywords - Deep Learning, Internet of Things (IoT), Convolutional Neural Networks (CNN)