Publication Details
Overview
 
 
Jelle Hofman, , Xuening Qin, Esther Rodrigo, Wilfried Philips, Nikos Deligiannis, Valerio Panzica La Manna
 

Environmental Modelling and Software

Contribution To Journal

Abstract 

Air quality monitoring in heterogeneous cities is challenging because a high resolution in both space and time is required to accurately assess population exposure. As regulatory monitoring networks are sparse due to high investment and maintenance costs, recent advances in sensor and IoT tech- nologies have resulted in innovative sensing approaches like mobile sensing to increase the spatial monitoring resolution. An example of such an opportunistic mobile monitoring network is “Snuffelfiets”, a project where air quality data is collected from mobile sensors attached to bicycles in Utrecht (NL). The collected data results in a sparse spatiotemporal matrix of measurements which can be completed using data-driven techniques. This work presents validation results of two ma- chine-learning models to infer air quality sensor data in both space and time a deep learning model based on Variational Graph Autoencoders (AVGAE) and a Geographical Random Forest model (GRF). The models are applied on different mobile datasets with subsequent temporal validation exercises at available regulatory monitoring stations following the FAIRMODE modelling quality objectives protocol. This work demonstrates the potential of data-driven techniques for spatiotem- poral air quality inference of sensor data as the considered models perform well in terms of accuracy and correlation. Moreover, the scalability of the approach is demonstrated as the models perform on different pollutants (PM2.5 and NO2) and mobile platforms (road vehicles and bicycles). The observed performance metrics approach current state-of-the-art physical models in terms of perfor- mance while needing much lower resources, computational power, infrastructure and processing time.

Reference 
 
 
DOI scopus