An IoT based Pollution Monitoring system using Data Analytics Approach

Abstract

Air pollution occurs when harmful gases such as CO, NH3, etc. concentration levels increase above the threshold level specified by the World Health Organization (WHO). Among this, one of the very important parameters is particulate matter. These are tiny particulate that they reach directly to the lungs and cause breathing problems. The standard level of range for pollution is already given by the central governing body of India i.e., Central Pollution Control Board (CPCB) in terms of the Air Quality Index (AQI). In this paper, a system for detecting the air pollution index with the help of Raspberry Pi-based system on IoT Technology which sends an Emergency Notification (EN) if there are any chances that the air pollution may raise above the given threshold in the future is developed which measures physical parameters like Temperature, Humidity, Dew point, Wind Speed and pollutants parameters like Suspended Particulate Matter (SPM) and Carbon Monoxide (CO) are monitored, and the effect of these parameters in pollution level is being predicted for pollution monitoring. The main objective of this is to apply the Machine Learning Algorithm for the prediction and analysis of gas sensors concentration levels, the effect of physical environmental parameters so that we can analyse the future concentration levels (AQI) level of the gaseous pollutant and based on this an Emergency Notification (EN) is send to the public as well as the concerning authorities. A system is developed for monitoring and alerting in real-time. We are discussing the different methods used in Machine Learning Algorithm i.e., Support Vector Machine (SVM) and Random Decision Forests (RDF) to predict the multivariate time series for forecasting and to use these predicted values to send an Emergency Notification (EN).