Design of IoT-based Air Quality Meter Module and Air Quality Analysis with Machine Learning

Abstract

This study proposes an ARM based air quality module placed to public transport vehicles for analyzing the effect of PM2.5 and PM10 particles in the cities in real-time using Internet of Things. The STM32 microcontroller is used for obtaining the data from the PM, humidity, and temperature sensors. The data collected from the sensors are sent to the i.MX6UL microprocessor using RS-485 connected to the internet portal with an Ethernet module. The microprocessor sends the data to the Microsoft Azure Hub in-on-line, and it is also recorded via the computer. The obtained data is analyzed for air quality-meteorological variables and the regression models are implemented via machine learning algorithms. PM2.5, PM10, humidity and temperature data are evaluated with R2 test and root mean square error for regression models. The Random Forest algorithm shows better results among other used regression models.