Regression based estimation of traffic flows using multi-source data

Francesco Bianchi, Giovanni Felici, Giuseppe Stecca

Keywords: regression analytics, digital twin of mobility, multi-source data

Abstract:

The successful exploitation of emergent digital technologies and advanced location and sensor-based data analytics can effectively
support urban mobility. In this context, the digital twin of mobility framework considers the real-time availability of data from diverse sources. The integration of these sources enables the simulation of mobility behavior and facilitates decision support at the urban scale.
In this work we propose a procedure to integrate multi-source data to road network arcs with the aim to calibrate traffic models.
Data are gathered from different types of road sensors but also from cellular network. Presence data are allocated to the road arcs. Regression models are run on the collected dataset and results allow to infer relations among static presence data and traffic flow patterns. A test case on the Italian city of Matera is presented.

Submission Number: 36