Implementation of Smart and Selective Gas Sensor System Empowered using Machine Learning over IoT Platform

Abstract

With the help of an internal sensor composed of hollow spheres of tin oxide coated in platinum, this stand-alone, selective gas sensor system offers wireless monitoring and Internet connectivity. At different concentrations, it reliably detects volatile organic compounds (VOCs). Real-time VOC identification is made possible using a machine learning model that has a 96.43% accuracy rate and a quick prediction speed of 310 µs. A low-power microcontroller and Bluetooth are used to connect the system, and real-time data can be accessed via an Android app or webpage that leverages cloud services. The system is validated and tested, enabling remote, self-sufficient applications that progress the Internet of Things' gas sensing capabilities.