IoT Network for PM Monitoring: Development and Deployment

Abstract

Monitoring and analyzing air quality is a challenging task without understanding influencing factors. Conventional air pollution monitoring is limited to few locations and accessing data is difficult. Internet of Things (IoT) provides a solution by enabling cost-efficient networks of particulate matter (PM) monitoring devices that are easily connected to the internet. PM monitoring portable sensors combined with IoT offer a more efficient and widespread PM monitoring solution, overcoming issues with bulky and expensive traditional systems. This thesis focuses on development and deployment of PM monitoring device and the establishment of a dense PM monitoring network. This thesis focuses mainly on four aspects. In the first aspect, the development of an end-to-end low-cost IoT system for a densely deployed PM monitoring network carried out an extensive field study in a region of the Indian metropolitan city of Hyderabad. A total of 49 devices were deployed in an area of 4 km2 , with 43 of them developed specifically for this study and the remaining six obtained from external sources, a density never realized in any metropolitan city worldwide and mostly in developing countries like India in the past. A total of 15 devices were connected to Wi-Fi, primarily within the campus area and wherever Wi-Fi connectivity was available, whereas 34 devices were equipped with 2G eSIM. The low-cost sensors were carefully calibrated to account for seasonal variations by utilizing a highly precise reference sensor. Also, a robust device was made that can cache data to avoid loss due to communication outages. The second aspect of this thesis focuses on gathering a significant quantity of data, nearly 20.7 million. This unique dataset offers great opportunities for future research and is examined to evaluate whether a concentrated deployment of PM monitoring devices was necessary. Third, various analytical methods, such as mean, variance, spatial interpolation, and correlation, were utilized to produce informative findings about the fluctuations of PM over time and across seasons. In order to understand the impact of firecracker detonations during Diwali festival evenings, a spatio-temporal analysis of PM values was conducted. As part of our analysis, we also tried to answer an important question concerning decision-makers about the optimum density required for effectively monitoring street-level pollution. For the considered scenario, we demonstrated that PM monitoring devices should be deployed at most 350 m apart to accurately capture the spatial variability of PM. Finally, this thesis examines various challenges encountered during the pre-deployment and postdeployment of PM monitoring devices. It addresses issues such as the design of low-cost devices, pre-deployment calibration, and seasonal calibration. Also explores challenges related to power supply, theft, environmental factors affecting sensor performance, and hardware failures. Additionally, the study investigates hardware reset and corrupt or redundant data. This thesis examines the challenges that arise during the deployment of PM monitoring devices and to propose feasible solutions to mitigate these issues, thereby offering valuable insights into the effective implementation of PM monitoring technology. Overall, the thesis highlights the importance of dense deployment and addresses the challenges to ensure reliable and accurate data collection.