Urban Traffic Simulation Using Mobility Patterns Synthesized from Real Sensors

Abstract

Smart cities are an ongoing research topic with multiple sub-research areas, from traffic control to optimization and even safety. However, testing the new methodologies or technologies directly in the real world is an almost impossible feat that, inclusively, can result in disaster. Thus, there is the importance of simulation. Simulation enables testing new and complex methodologies and gauging their impact in a realistic context without adding any safety issues. Additionally, these can accurately map real-world conditions depending on the simulation configuration. One key aspect of the simulation is the traffic flows in the simulated region. These may be hard to find and, if ill-set, may introduce bias in the results. This work is on the characterization of the traffic in the city center of Guimarães, Portugal. An urban simulation scenario was established, using SUMO as the mobility traffic simulator, with traffic patterns derived from real-world data provided by Guimarães City Hall and using Eclipse MOSAIC for extended vehicular simulation. Apart from mobility patterns analysis, this work also provides publicly accessible datasets, simulations, and applications made available to future research works.