Monitoring and Prediction of Air Quality System using Internet of Things (IoT)

Mohammed Saad Ashraf Alrubaye; Azizan As’arry; Muhammed Amin Azman; Mohd Zuhri Mohamed Yusoff; Khairil Anas Md Rezali; Ali Zolfagharian

doi:10.37934/araset.48.1.6176

Authors

Mohammed Saad Ashraf Alrubaye Department of Mechanical & Manufacturing, Faculty of Engineering, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia
Azizan As’arry Department of Mechanical & Manufacturing, Faculty of Engineering, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia
Muhammed Amin Azman Department of Mechanical & Manufacturing, Faculty of Engineering, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia
Mohd Zuhri Mohamed Yusoff Department of Mechanical & Manufacturing, Faculty of Engineering, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia
Khairil Anas Md Rezali Department of Mechanical & Manufacturing, Faculty of Engineering, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia
Ali Zolfagharian School of Engineering, Deakin University, Geelong 3216, Australia

DOI:

https://doi.org/10.37934/araset.48.1.6176

Keywords:

Monitoring system, Air quality, Prediction and regression models, Internet of Things (IoT), ML.NET algorithm

Abstract

Air pollution is a global issue causing 7 million deaths annually, primarily due to dangerous compounds and inhalable particles. The World Health Organization estimates that air pollution is a significant threat to human health and safety. The Internet of Things (IoT) is utilized in this research to analyse sensor data from various environmental monitoring devices. The objective is to build an IoT-based air quality system that assesses air quality conditions in specific areas and analyses air levels of various compounds. The study employs IoT technology to analyse data from various environmental sensors, aiming to create an IoT-based air quality system. The sensors include measuring NH, C6H6, VOCs (MQ135), CH4 (MQ5), and particle matter (PM2.5). The ESP32 microcontroller and thinger.io platform is used to develop the system. FastTree and Generalized Additive Model (GAM) are machine learning methods applied to predict and analyse air quality data. FastTree utilizes gradient-boosting to enhance accuracy, while GAM employs smooth parts, represented by splines, for relationship modelling. Evaluation metrics R2, RMSE, MSE, and MAE assess model performance, for MQ135 with FastTree outperforming GAM with an R2 of 91.91% compared to 78.05%. The IoT-based air quality system is user-friendly and effective, tracking harmful components against thresholds. Future work includes adding cameras and printed circuit boards for expanded analysis.