An IoT-Based Air Quality Monitoring with Deep Learning Model System

Abstract

Air pollution occurs when the concentration levels of environmental gases including CO2, NH3, etc., go above the optimum level. As the AQI is being calculated and as per the Central Pollution Control Board (CPCB), there is a standard level of ranges for pollution level. This paper presents monitoring of the pollution level using Raspberry Pi 3 based on IoT technology. Here, the temperature, humidity, dew point and wind speed parameters are also monitored and these parameters are used as datasets for prediction of pollution forecasting. Then, the target of this project is applying the deep learning concept for the prediction and analysis of gas sensors' pollution level so that we can analyze the pollution level due to the pollutant gases based on prediction analysis. Various experiments were performed for the validation of the development of the system for real-time monitoring. Here, we are discussing the different methods used in deep learning, i.e., artificial neural networks (ANN), multilayer perceptron (MLP) and recurrent neural networks (RNN), using LSTM model to analyze and predict the multivariate time-series forecasting.KeywordInternet of ThingsRaspberry Pi 3Recurrent neural networks (RNN)Long short-term memory (LSTM)Air Quality Index (AQI)