Smart Environment: AI-Driven Predictions and Forecasting of Air Quality

Abstract

Addressing the critical issue of air quality in the Coimbatore region, this study introduces a novel approach for continuous monitoring and forecasting of air pollution. By utilizing the Internet of Things (IoT) technology integrated with Artificial Intelligence (AI) methods, this research focuses on monitoring and forecasting three major pollutants such as Ozone (O3), Ammonia (NH3), and Carbon Monoxide (CO). The proposed IoT-based sensor nodes collect the real-time data and give the resultant data as an input to the Naive Bayes (NB) for classification and Auto-Regression Integrating Moving Average (ARIMA) for optimization. The optimized model parameters are obtained and then validated by using performance metrics like Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). Deploying a machine learning algorithm on a Raspberry Pi-3, the proposed system ensures efficient monitoring and forecasting of air pollutants 24/7 through an online open-source dashboard.