Crowd Guidance in Building Emergencies: Using Virtual Reality Experiments to Confirm Macroscopic Mathematical Modeling of Psychological Variables

A general challenge during a building emergency evacuation is guiding crowd to the best exits, given potential hazards and blockages due to high density use. Although computer simulation programs such as FDS+Evac allow researchers to evaluate various guidance policies under different circumstances, computational complexity limits their use during an actual emergency. A second limitation of such programs currently available is that they can only model certain psychological variables that affect evacuation. We suggest two innovations to address these difficulties. First, using macroscopic models, mathematical techniques can allow for rapid optimization of guidance that could eventually be used to provide real-time use during emergencies. Second, we conduct virtual reality experiments using human participants to provide confirmation of our models, and offer insights into how psychological factors not yet available in FDS+Evac will affect evacuation outcomes. Results of an initial VR experiment are presented.