Generating Synthetic Passenger Data through Joint Traffic-Passenger Modeling and Simulation

Real passenger data available to city planners are usually incomplete. The goal of our work is to generate synthetic passenger data using a novel methodology that leverages joint traffic-passenger modeling and simulation on a city scale. A demonstration of such an idea in generating synthetic bus passenger data was implemented. Specifically, we 1) learned a bus passenger demand model from indirect people-mobility data to generate bus passenger demand samples, and we 2) developed a bus passenger behavior model, which runs jointly with a traffic simulator (SUMO), to generate synthetic bus passenger data. We applied the proposed methodology for a case study of Porto city, Portugal. The synthetic bus passenger data presents significant similarity in terms of spatial-temporal distributions to the real-world bus passenger data collected by the bus automated fare collection (AFC) system in Porto.