An evacuation model considering the phototactic behavior of panic pedestrians under limited visual field

In the emergency evacuation process, pedestrians affected by their panic will have evacuation behavior that rushes toward the light. To study the impact of panic pedestrian phototactic behavior on the evacuation process, a multi-model coupled cellular automata model is proposed. First, a heterogeneous model of emotional contagion based on the OCEAN model and the SIS model is established, which considers the influence of individual characteristics such as personality, gender, and age on emotional contagion, and the increase and decay of pedestrian emotion are also correlated with their location. Next, the "guidance field", "orienteering field", "herding field"and "phototactic field"are introduced to build an extended cellular automata model, pedestrians in different areas have different movement modes, and pedestrians with different emotional states in the same area also differ. Finally, the phototactic behavior of pedestrians in evacuation is reproduced by simulation, and the effects of factors such as emotional threshold, information transfer, visual field distance, light source position, and multiple light sources interference on the evacuation process are discussed. The results show that whether phototactic behavior is favorable or unfavorable to evacuation depends on the position of the light source. When the light source is located at the exit can guide pedestrians to escape, while located in other places will trigger local congestion. Increases in the emotional threshold, information transfer, and visual field distance can effectively mitigate the adverse effects of phototactic behavior. Additionally, the light sources in other positions will cause some interference with the light sources at the exit when multiple light sources exist, prolonging the evacuation time. Our research can provide some guidance for designing building structures and developing evacuation strategies in emergency situations. (c) 2023 Elsevier B.V. All rights reserved.