Air Quality Monitoring and Prediction System Using Machine-to-Machine Platform

This paper presents an ambient air quality monitoring and prediction system. The system consists of several distributed monitoring stations that communicate wirelessly to a backend server using machine-to-machine communication protocol. Each station is equipped with gas- eous and meteorological sensors as well as data logging and wireless communication capabilities. The backend server collects real time data from the stations and converts it into information delivered to users through web portals and mobile applications. In addition to manipulating the real time information, the system is able to predict futuristic concentration values of gases by applying artificial neural networks trained by historical and collected data by the system. The system has been implemented and four solar-powered stations have been deployed over an area of 1 km2. Data over four months has been collected and artificial neural networks have been trained to predict the average values of the next hour and the next eight hours. The results show very accurate prediction.