Virtual experimental study on pedestrian evacuation dynamics in view-limited environment

In emergency situations, factors such as smoke and dimly-lit conditions that lead to limited visibility for pedestrians often become an important reason for preventing more efficient evacuation. Considering factors such as cost, preparation time, and especially the safety and ethical issues of the participants, it is difficult to organize evacuation experiments with limited vision in reality. Therefore, in this paper, a multi-person online virtual experiment platform, including drawing, motion, and communication modules, is developed using the high security and expandability of virtual experiment technology to conduct such investigations that are difficult to organize safely and efficiently in reality. Firstly, the validation experiments were organized to compare with the control experiments regarding route choice and motion characteristics. The results showed that the platform could reproduce not only fundamental diagram but also route choice features of the real-life evacuation experiments. Secondly, the motion characteristics and following behavior of pedestrians during evacuation with different number and location settings of exit and different visual fields with limited vision were further studied. The results show that the movement characteristics and following behavior of pedestrians with limited vision are related to social information differences due to the visual field range (VFR) and building environment factors such as the number and location of exits. With VFR increases, the pedestrian evacuation time and movement area decrease and remain stable after VFR reaches 3.0 m. The average collision number of pedestrians increases with VFR, and the curve of the number of collisions in the dual-exit scenario is “V” shaped. The wayfinding time of pedestrians decreases with the VFR increase, and the blocking time increases accordingly. As the increase of VFR, the movement consistency and the following frequency of the crowd is higher; the larger of VFR in the single-exit scenario, the higher maximum frequency value that pedestrians are more likely to engage in following behavior, while they tend to make more autonomous decisions in the dual-exit scenario.