Eqasim simulation using mobile phone signalling data: A case study of Shanghai, China

Abstract

eal data-driven traffic simulation provides analytical support for the assessment of Automated Vehicle introduction. This study differs from the MATSim simulations combined with mode choice modules used in previous studies, instead using eqasim pipeline simulations. Reliable travel demand data is generated based on a detailed analysis of the mobile phone signalling data. Combing mobile phone signalling data and other source data, the city’s travel characteristics are identified by analyzing the spatial-temporal travel flow in Shanghai from the perspectives of regions and users. Subsequently, the data containing persons’/agents’ activity patterns and travel modes is used to obtain the travel demand considering the four main travel modes, which are respectively car, public transport, bike, and walk. Given the city scale, the data within the outer ring of Shanghai is selected for the traffic simulation. Multinomial probit model (MNP) is used to simulate the agent’s mode choice decisions in this study. Its parameters are calibrated against mode share and trip distance. We tested the convergence results of the scenario through experiments and performed a comparative analysis with the reported travel survey data. This study can not only provide preliminary guidance for eqasim simulation when using real data, but also shed some lights on the application of mode choice and analysis involving AVs in the mega-cities of the near future.