Effects of risk information on pedestrian evacuation during fire emergencies: Virtual experiments and survey

Abstract

In fire emergencies, pedestrians often face uncertainty about the safety of evacuation routes due to a lack of risk information, which makes the emergency evacuation highly challenging. Although risk information is crucial for effective evacuation, the effect of varying levels of risk information on fire evacuation remains insufficiently explored. This study aims to address this gap by investigating the influence of different levels of risk information on pedestrian evacuation through virtual experiments and survey. We conducted a series of multiplayer fire evacuation experiments in Minecraft. The risk information provided to pedestrians before evacuation was controlled at three levels: no risk information, partial risk information, and complete risk information, by giving different risk information cues. The results of experiments indicated that risk information can improve pedestrian evacuation efficiency during fire emergencies, while more complete risk information can further improve pedestrian evacuation efficiency. Additionally, an information transfer phenomenon was observed during fire evacuations, in which pedestrians adjusted their behaviour based on perceived fire, leading others to avoid the fire in advance. The analysis of pedestrian behaviour revealed that pedestrians did not exhibit imitative behaviour in evacuation direction choices during fire emergencies. Pedestrian route choices were significantly influenced by the initial distance of pedestrians relative to the exits. Moreover, complete risk information cues did not significantly influence pedestrians’ route choices compared with partial risk information cues. A post-experimental survey was conducted to assess pedestrians’ route choice preferences under different risk information cue scenarios. The results showed that pedestrians preferred the nearest route.