Enhancing pedestrian flow realism at crossings: a comfort zone projection-based social force model

Accurately simulating pedestrian behavior is crucial for traffic planning, facility design, and safety evaluation. However, existing social force models have limitations in capturing microscopic pedestrian behaviors, especially regarding bidirectional avoidance and path conflicts. To overcome these shortcomings, this work proposes an enhanced social force model based on the projection of pedestrian comfort zones. The core innovation lies in quantifying conflict intensity via projection overlap, incorporating refined local interaction mechanisms to enable intelligent overtaking, following, and avoidance behaviors. Experimental results demonstrate that the improved model effectively mitigates unreasonable weaving and oscillations, exhibiting strong self-organizing properties in bidirectional pedestrian flows. Moreover, by introducing a custom-defined pedestrian conflict index, the model offers deeper insights into macroscopic flow characteristics. Simulation results indicate that the projected social force model (PSFM) allows pedestrians to complete crossings with minimal speed loss and fewer movement conflicts compared with both the classic social force model and an existing improved version. These findings offer valuable insights into pedestrian behavior and contribute to the development of more realistic traffic simulation systems.