An IoT-Based Air Pollution Monitoring System for Smart City

In this paper, we present the results of research on the design and development of a gas analytical unit, an electronic nose, with an autonomous power supply, as well as the function of wireless data transmission via the Long-Range Wide Area Networks protocol. We built the device on a circuit board featuring eight gas sensors (MQ series), an STM32F103C8T6 microcontroller with 12-bit analog-to-digital conversion, a RAK3172 wireless data transmission module, voltage dividers, and a USB serial adapter. It also has an HDC2080DMBR temperature and humidity sensor. We evaluated the electronic nose performance towards standard mixtures of carbon dioxide, carbon monoxide, nitrogen dioxide, ammonia, and hydrogen sulfide, which are important analytes in outdoor and indoor air quality monitoring. Multidimensional linear regression methods were employed to build calibration curves for electronic nose sensors, which allowed for converting the measured sensor responses into quantitative values of the concentrations of the analyzed substances. The estimated coefficients of determination, R2, approach 1 indicating a strong correlation between the approximating model and the original data. The multivariate chemiresistive responses were successfully clustered using the k-means algorithm, revealing clear distinctions between sensor data associated with different gas concentrations. The k-NN classifier, applied to the clustered data, achieved 100% accuracy in identifying gas concentrations, underscoring the potential of the proposed methodology for gas sensing applications. We proposed and implemented the architecture of a network of mobile gas analyzers as the Internet of Things devices for environmental monitoring using the VK Cloud Platform.