An improved social force model for pedestrian-vehicle interaction simulation
Date:
Zhou, Z., Liang, H., & Ye, Y.* (2025, July 22-25). An improved social force model for pedestrian-vehicle interaction simulation [Poster Presentation]. The 25th COTA International Conference of Transportation Professionals (CICTP2025), Guangzhou, China.
Abstract: Pedestrian simulation is crucial in urban planning, safety assessment, emergency preparedness, traffic management, and the advancement of autonomous vehicles. The social force model effectively simulates pedestrian fundamental movement behavior; however, it still has limitations in handling dynamic pedestrian-vehicle interactions in complex traffic scenarios. Therefore, this paper proposes an improved social force model to enhance the simulation of pedestrian-vehicle interactions. Accordingly, a decision-making module is incorporated into the social force model, dynamically simulating pedestrian decision-making behavior during interactions by integrating pedestrian and vehicle speed and positional information. Additionally, a velocity difference factor is introduced to more accurately capture the impact of varying vehicle speeds on pedestrian behavior, resulting in a more realistic representation of pedestrian reactions in diverse traffic conditions. The model’s effectiveness is validated through qualitative and quantitative comparisons with real-world pedestrian trajectory data. This work sheds light on the development of more advanced pedestrian-vehicle interaction models, as the results demonstrate that the improved model significantly enhances the accuracy of capturing pedestrian responses under different vehicle speed conditions, overcoming the limitations of traditional models. This study notably improves the simulation accuracy of the social force model in pedestrian-vehicle interaction contexts and provides a more reliable tool for simulating pedestrian behavior in complex traffic environments.