Calibration of Low-Cost Particle Sensors by Using Machine-Learning Method

Abstract

The measurement of particle matter (PM) of mass concentration by low-cost PM sensor is strongly influenced by environmental factors such as humidity, temperature, wind speed, wind direction. In this study, we developed a machine learning-based calibration method for low-cost light-scattering PM sensor. A Feedforward Neural Network (FNN) was used to compensate for the effect of environmental factors on the PM measurements. Experimental data were collected from 20 March - 6 May 2018 in central Taiwan, and used to train and evaluate the calibration model. Before calibrating PM sensor, the PM2.5 mass concentration of low-cost PM sensors have the lowest values of R-squared (R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">)</sub> , with 0.618±0.033 as compared to the Environmental Protection Agency (EPA) approved federal equivalent method (FEM) instrument (BAM-1020, Met One Instruments). After calibrating PM sensor by using the FNN calibration model, the PM2.5 mass concentration of low-cost PM sensors show the highest linearity with an R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> value of 0.905±0.013 for BAM-1020. It demonstrated that the machine-learning method could be used to calibrate a low-cost PM sensor and improve its accuracy.