Simulation of Running Crowd Dynamics: Potential-Based Cellular Automata Model

Abstract

Running is often accompanied by fear or panic during emergency evacuation, and the evacuation of pedestrian crowds at different speeds poses additional challenges in terms of modeling and simulation. To investigate the fundamental interference of running pedestrians in the evacuation process, in this study, an extended cellular automata model was developed to simulate the evacuation of mixed groups of walking and running pedestrians in an area with multiple exits. The innovations of this extended model are the application of the dynamic potential field algorithm considering panic propagation, pedestrian running caused by panic, and pedestrian ratio and obstacle layouts. Running pedestrians, converted from walking pedestrians, were recognized by the number of k-nearest neighbors in the moving direction based on the Manhattan distance method. The effects of initial pedestrian density and obstacle layout were studied using numerical simulations. The simulation results indicate that a certain number of running pedestrians are needed to improve the evacuation efficiency, and the panic is contagious to others in that walking pedestrians transform into running pedestrians to accelerate the evacuation speed. The research provides insights for improving pedestrian-evacuation efficiency in facilities similar to the scene used in the experiment.

Keywords

• Running pedestrian flow
• cellular automata model
• pedestrian conversion
• K-nearest neighbors method
• dynamic potential field