Modeling and Simulation of Exit Selection Behavior in Pedestrian Evacuation Based on Information Perception and Transmission

This paper aims to present an improved evacuation model, which is capable of simulating individual exit selection behavior based on the acquisition and processing of information, especially in dangerous and unfamiliar environments. Firstly, an evacuation model was improved by the introduction of a floor field of gas concentration and an exit selection model, considering the congestion avoidance and danger avoidance behavior. Secondly, the process of information perception and transmission was studied and introduced into the model with a set of rules. Finally, real experiments in a simple double-exit room were conducted for model validation and parameter setting, and simulation experiments in scenarios with an unknown hazard or unknown exits were conducted to confirm the necessity and rationality of introducing information perception and transmission. The simulation results show that, with the increase in perception distance or trust extent, the pedestrian safety increases. The critical values of perception distance or trust extent, below which some people cannot acquire any new information, vary depending on the pedestrian density. When the density is high, the influence of perception distance or trust extent reduces, and the probability of reselecting an exit increases, which causes the safety of pedestrians to decrease.