Digital Twin for mobility: simulation integrated approach for demand matrices calibration with empirical data

A Digital Twin for Mobility (DTmob) is based on the digital representation of the real-world transport system. A crucial element underlying the development of such systems is an accurate simulation of traffic flows within the road network, which is generally based on transport demand data in the form of Origin-Destination (O-D) matrices. Estimating demand matrices accurately is a complex challenge, as they are hindered by incomplete data, inaccuracies in transport models and temporal-spatial variability. In this paper we present an integrated methodological approach for the calibration of O-D matrices via the implementation of an iterative calibration procedure incorporating a traffic survey dataset, followed by the assignment procedure performed by a macrosimulation model. This study addresses this issue by employing Nielsen’s Single Path Matrix Estimation Method (SPME) jointly with the macrosimulation tool PTV Visum, developing a script in Python to achieve their integration. We tested it on the real-case study of Catania urban road network (Italy). Several scenarios have been analyzed and evaluated through the GEH statistic as KPI, also providing results with Ordinary Least Square (OLS) method as benchmark. Obtained results prove the validity of our integrated framework as the threshold in terms of the difference between estimated and counted flows was satisfied in all performed analyses. Research findings lay the basis for investigate further calibration algorithms and for exploring the use of real-time data in line with DTmob requirements.