We present a case study on synthesizing realistic vehicle mobility using link traffic information extracted from
surveillance videos. Our trial has been conducted in a town of Toyooka city, Japan, where many people visit using private vehicles in tourism seasons, which causes parking lot issues in the center of the town. We have developed an automatic vehicle tracking system that can measure the link traffic in the captured videos. Using the information and a set of potential routes between origin and destination in the town, we estimate, for each route, the volume of the traffic flow on the route. Finally, we generate micro-level vehicle mobility data using a traffic simulator, minimizing the Mean Absolute Percentage Error(MAPE) between the measured and generated link volumes. The synthesized mobility will be used for improving the traffic, for example, the adoption of new traffic regulations in the region.
The reproduced mobility data obtained in this study are available here.
Published Paper
- Kazuki Hayashi, Akihito Hiromori and Hirozumi Yamaguchi (Osaka University, Japan); Masaki Suzuki and Takeshi Kitahara (KDDI Research, Inc., Japan, "Synthesizing Town-scale Vehicle Mobility from Traffic Surveillance Cameras: A Case Study", Proceedings of the 2022 IEEE International Workshop on Pervasive Computing for Vehicular Systems Co-located with IEEE PerCom 2022, pp. 593-598