Reliable Low-Cost Air Quality Monitoring Using Off-The-Shelf Sensors and Statistical Calibration

Abstract

Modern cities are densely populated spaces and number of people living in cities is increasing rapidly by years. The air monitoring stations exist in most of the cities to monitor air pollution. However, their number is insufficient having in mind the high cost of stations, as well as annual calibration cost. The potential solution is to use low-cost off-the-shelf sensors to monitor related air quality parameters, but they are not reliable due to the low accuracy, calibration issues, and short life cycle. In this paper, the methodology is proposed for calibration off-the-shelf air quality sensors using statistical algorithms and offset values from the official public measurement stations. The possibilities are analysed to improve the reliability of low-cost sensors by processing the obtained raw data. Special attention is devoted to the detection and elimination of short intervals when the raw results have the extraordinary high value-peaks and to the corresponding interpolation of these data. New algorithm for “peaks” detection and elimination is proposed and evaluated. Common Air Quality Index (CAQI) is calculated and evaluated in comparison with public monitoring station. It is shown that low-cost sensors could be used with high reliability as a complementary network to public monitoring stations.