Ambient particulate matter monitoring system using SDS011 sensor utilizing machine learning approach and ambit of blockchain technology

Abstract

In recent days, air contaminated with Particulate Matter Pollutants (PM2.5and PM10) plays a major role in affecting the quality of air and it has a strong effect on the environment. PM2.5and PM10 released in the air due to the huge increase of the manufacturing yards have led to significantly increased and severe health issues for all living beings around the globe. These Particulate Matters emitted from the industrial units can stay suspended in the air for a very prolonged time. Thus, it is essential to measure and monitor the ambient Particulate Matter emission rate near the industries. However, the existing fixed air quality monitoring stations are challenging, nevertheless, due to thehigh energy utilization, small-scale monitoring, and privacy concerns. To resolve this concern, a low-cost portable sensor NOVA SDS011 can be mounted near the industries to study the impact of the concentration level of the Particulate Matter Pollutants (PM2.5and PM10) released from those industries. The entire detected values are transmitted to a scalable IoT platform in the cloud, which allows the processing, analyzing, and exportation of data as.xls/.csv files. These exported data are pushed through various Machine Learning (ML) Classifier algorithms to classify the data into multiple classes based on its AQI (Air Quality Index) values. In this study, various ML classifier algorithms were used and analyzed in which Random Forest Classifier Algorithm had performed with utmost accuracy to classify the six different classes of AQI. This study also explains the limitation of privacy concerns in existing systems and the scope of using blockchain technology to overcome security concerns.