Human factors analysis and behavior modeling for the simulation of evacuation scenarios

Publication from Digital

Wagner, V. and Kallus, K.W. and Neuhuber, N.J. and Schwarz, M. and Schrom-Feiertag, H. and Stubenschrott, M. and Pszeida, M. and Ladstaetter, S., Paletta, L.

Ergonomics and Human Factors in Safety Management , 1/2016


In the case of evacuation events, human factors play an important role for the effective outcome of evacuations. Therefore, the design and organization of evacuation systems can be seen as crucial factors. As theoretical framework the theory of behavioral inhibition and activation systems can be considered to describe behavioral aspects of human beings in critical situations. Furthermore, the behavior of evacuation assistants can be seen as an additional factor that may influence the behavior of evacuees. Within an experimental study (Wagner et al., 2015) a multilevel approach was used to combine assessments of the subjective emotional states, behavior, and objective psychophysiological responses of the study participants. The results showed that the occurrence of an evacuation assistant influences the behavior and the emotional state of evacuees while acting in different conflict situations. Overall, the results give important indications to improve evacuation situations by avoiding critical situations in which persons tend to become unconfident and therefore become incapable of action. From the correlation between the multisensory perceptual and psychophysiological data and situational context a visual perception model was developed and a behavior rule set that parameterizes an agent based simulation model was extracted. The perception model and the behavioral model was integrated in an agent based simulation model to allow an evacuation analysis reflecting the observed human social behaviors, through simulating the cognitive processes of individual agents and interactions among neighboring agents in the simulation environment. For the evaluation of the simulation the experimental study was modeled in the simulation environment and the results were compared to the empirical observations. The developed autonomous agent was capable to reproduce empirically observed human behavior and enables to simulate scenarios with a high degree of realism. As a result the simulation supports improved placement of guidance components, impact assessment of evacuation assistants, capacity analysis of escape routes, the identification of critical areas, and the duration of evacuation.