An extended model describing pedestrian evacuation considering pedestrian crowding and stampede behavior

Abstract

In recent years, stampede accidents have occurred frequently, resulting in significant casualties and property losses. In this paper, an extended cellular automata model is designed to investigate the behavior of pedestrian stampedes. The concept of pedestrian crowding, support, and friction is introduced and viewed as an influencing factor of normal pedestrian falls Besides, the state of pedestrians after falling to the ground is analyzed. The fallen pedestrians may return to normal or be trampled to death. Through the numerical simulation, the stampede phenomenon in pedestrian evacuation in public places is explored, and the influences of pedestrian density, pedestrian distribution mode, exit setting, and obstacles at the exit on the evacuation results are discussed. The analysis demonstrates that the greater the pedestrian density, the more the pedestrians fall to the ground and die. Moreover, pedestrians are more likely to die after falling to the ground with the increase in pedestrian density. At the beginning of the evacuation, the risk of the pedestrian stampede in decentralized distribution is less than that in centralized distribution. The exit is located on the left or right of the center of the wall when the pedestrian density is small, making it less dangerous than the middle position. The safety is higher when the number of exits is large. Additionally, the existence of obstacles at the exit can reduce the risk of stampede accidents when the pedestrian density is small. Our research reveals the mechanism and influencing factors of the trampling process to a certain extent, playing a reference and guiding role for future research.