Virtual Reality and Human Behaviour in Fire: An Exit Choice Experiment to Measure Social Influence and Role-Rule Influence

In the event of a fire emergency in a built environment, human behaviour plays an important role, since building occupants need to make several evacuation decisions. Exit and route choices are one of them. It is fundamental to investigate which factors influence this choice. Understanding how humans weigh different factors when they make a choice is crucial to design a safe built environment for evacuations. The main objective of this work is to study the exit choice of occupants during fires and develop an exit choice model which could be implemented in future evacuation software tools. In particular, this research focuses on the influence of two different factors: the social influence and the role-rule factor. To achieve this goal, this work uses a new immersive virtual reality experiment to investigate exit choice during a fire evacuation in a metro station. This experiment involves 131 random participants who were asked to choose between two different possible exits in different evacuation scenarios. These scenarios differ in the presence or not of an instructor guiding participants, the authority of the instructor and the presence or not of other evacuees who use a specific exit. Using the collected data, a discrete choice model was estimated to investigate if and how these factors affected the decisions of the participants. The results of the models show that all the factors considered influence the participants’ choices. In addition, the results show that the higher the authority level of the instructor the more participants follow the instructions. Finally, other models were developed to analyse the impact of demographic factors on the choice of participants. In this case, the results do not show evidence that demographic factors (i.e., age, nationality, handedness and BMI) have an impact on the exit choice. This work provides several novelties in terms of findings, and it has implications for both fire safety designs and future VR evacuation experiment designs. This study is the first one that investigates a unique combination of the variables mentioned above and their interactions.