A Modified Social Force Model for Pedestrian Behavior Simulation at Signalized Crosswalks

Abstract

In this study, a microscopic simulation model is developed for pedestrian behavior analysis at signalized crosswalks. It takes into account the special characteristics of pedestrian crossing behavior, such as group evasion with surrounding pedestrians, collision avoidance with conflicting vehicles, response to signal control, and response to crosswalk boundary. The classical social force model is modified to reproduce repulsive and attractive phenomena. Furthermore, a generic calibration approach is proposed based on maximum log-likelihood estimation, which enables to identify significant parameters in a statistical way. Last, a rigorous validation is conducted to confirm the model performance in terms of collision avoidance behavior with conflicting pedestrians and vehicles.