Spatial Calibration of Mobile Sensor Traffic Data on Road Networks

The estimation of traffic volumes on a street network represents a fundamental step in urban engineering and transport studies. Traditional methods for obtaining traffic data rely on manual counts from trained operators or sparse automatic sensors. However, in recent years the growing availability of mobile sensors has led to a surge in methods for inferring traffic volumes from geo‐referenced mobile devices. This paper proposes a spatial statistical calibration framework that combines accurate fixed‐location counts with extensive GPS‐based mobile data to estimate traffic flows, while accounting for the street network geometry and its directionality. Finally, starting from the calibrated estimates, we propose a procedure to assess the performances of a reduced grid of fixed‐location sensors and, using an heuristic procedure, identify the optimal subset. The methodology is illustrated using data collected in the city of Leeds (United Kingdom).