AI-Navigated Fog Computing based Portable Ambient Particulate Matter Sensor Calibration

Bull Trench Kilns, shortly referred as BTK are one of the major sources of emitting Particulate Matter (PM) pollutants in air, which is harmful to the environment, and it has the capability to remain suspended in the air for very long duration. As conventional fixed pollution monitoring stations are sparsely located, and Brick industries are generally located in remote locations, BTK's pollutant emission rate is not exactly been recorded/monitored. In this study, portable PM sensor along with a reference sensor are mounted near the Brick Kilns and the pollutant's concentration in air along with the atmospheric temperature and humidity are measured. Since PM sensors are sensitive to atmospheric factors, its readings are calibrated using Machine Learning (ML) techniques, with the help of data collected from reference sensor, to provide more reliable data. The proposed AI-navigated three phase architecture is used to calibrate the readings obtained from the sensor, where Linear Support Vector Regression (LSVR) obtained best results.