An Integrated Sensing and AI Framework for EMF and NO₂ Exposure Assessment in Urban Environments: A Ghent Case Study

Abstract

Urban environments are increasingly exposed to environmental stressors such as electromagnetic fields (EMF) and air pollution, challenging public health and infrastructure sustainability. The widespread deployment of 5G networks, wire-less communication systems, as well as industrial emissions and heavy traffic has increased the need for advanced sensing and modeling techniques to monitor and mitigate potential risks due to pollution. In this study, we present an integrated urban sensing framework that combines low-cost sensor networks, data fusion methodologies, and machine learning models to predict and forecast environmental exposure to EMF and nitrogen dioxide (N02) in Ghent, Belgium. We evaluate the performance of different sensing technologies, data preprocessing techniques, and statistical and deep learning-based models for data imputation and forecasting; addressing the challenges of data calibration and sparsity. Our findings demonstrate the effectiveness and need for improving exposure assessment by providing valuable insights for urban planning, policy-making, and sustainable development.