Improved social force model based on pedestrian collision avoidance behavior in counterflow

The social force model is a widely used microscopic model for reproducing pedestrian behaviors. However, the traditional social force model lack consideration for pedestrian collision avoidance behaviors, such as the significant impact on model accuracy in scenarios like pandemics. In this study, we introduce a novel avoidance force component to the social force model to capture pedestrian collision avoidance behavior within counterflow. We conducted bidirectional pedestrian control experiments during the pandemic, directly tracking pedestrian movement trajectories and spatiotemporal characteristics. Key parameters were extracted and used to calibrate the model based on collision avoidance scenarios among pedestrians. Furthermore, we compared the performance of our proposed model with previous model research and experiments by using individual crossing time and order parameter as evaluation parameters. These findings provide an efficient tool for future applications that can realistically model pedestrian counterflow under conditions where pedestrian collision avoidance behavior is significantly pronounced.