Investigating the Effectiveness of Station Staff Guidance in Underground Urban Railway Fire Emergencies: Virtual Reality Experiment

Exploring the critical roles of various factors on determining the wayfinding performance of passengers during a subway fire emergency is important. By developing a virtual reality environment of an underground urban railway station, an immersive virtual reality escape experiment was conducted with a head-mounted device to record the wayfinding performance of 1,151 participants. The experimental data include trajectory data of the participants in the escape task (e.g., time and three-dimensional coordinates) and survey data of personal information. Binary logistic regression, general linear regression, LightGBM, and other methods were employed to analyze the effects of station staffs’ guidance, emergency broadcast, signage, and other factors on wayfinding performance, with station staffs’ guidance being a factor rarely addressed in previous studies. The large sample size enabled the application of ensemble learning methods and facilitated fine-grained group difference analyses that are difficult to achieve in traditional small-sample evacuation experiments. Results show that emergency broadcast and the guidance of station staffs had a highly significant impact on evacuees’ route choice. Education exhibited a positive influence on both evacuees’ route choices and evacuation time. Furthermore, information overload was observed among evacuees under the influence of multiple information sources. The results of this study can not only serve as a theoretical basis for policymakers to understand the determinants of wayfinding performance, but also provide guidance for underground urban railway station or metro station evacuation plan in a fire emergency.