A Cost-Effective Air Quality Monitoring System for the Global South

Abstract

Air quality monitoring is crucial for mitigating the health risks associated with ambient air pollution, particularly in low- and middle-income countries, where resources for such systems are often limited. Traditional monitoring solutions are expensive, resulting in sparse data coverage and inequitable access to air quality information. This paper introduces the AI_r system, developed by the South African Consortium of Air Quality Monitoring (SACAQM), as a cost-effective air quality monitoring solution designed for the Global South. Leveraging Internet-of-Things (IoT) technology and low-cost sensors, the system establishes a dense network that provides high-resolution, real-time air quality data. The architecture includes clustered Wireless Sensor Networks (WSNs) Long-Term Evolution (LTE) communication, with data stored in a NoSQL database and accessed via an interactive dashboard. Calibration against established air quality systems, such as the South African Air Quality Information System (SAAQIS), ensures data accuracy and reliability. Additionally, the system integrates Artificial Intelligence (AI) techniques, including Graph Neural Networks (GNNs), to model and predict air quality trends. The system has been successfully piloted with sensor deployments in schools in Soweto, Johannesburg. The AI_r system aims to democratize access to critical air quality data, supporting public health initiatives and policy development to improve air quality, particularly for vulnerable populations.