Experimental And Modeling Study On Pedestrian Movement Under View-Limited Condition

With the rapid development of economy and society,a variety of buildings and facilities with complex structures have emerged.Meanwhile accompanying by an increase of pedestrian travel frequency and large public events,large scale crowds are more likely to gather in many public places such as schools,stadiums,metro stations or train stations etc.The movement of individuals is severely constrained under high density,which is easy to cause crowded stampede.The current studies are mainly focused on pedestrian evacuation under good visibility.However,this is often not the case during fire emergencies,as smoke or a failure of the electrical power supply reduces the orientation significantly and the visual field of pedestrians is affected to varying degrees.Therefore,in this paper,the movement characteristics and evacuation process under view-limited condition are investigated by using both experimental and modeling methods.To obtain pedestrian movement data under view-limited condition,a single-file movement experiment was firstly conducted.Based on the analysis of experimental video and pedestrian’s trajectory,it is found that pedestrians prefer to follow the wall boundary and people in the front;With the increment of pedestrian density,the lateral sway amplitude of trajectory also increases;With the decrease of visibility,the phenomenon of stop and go at high density starts to dominate the traffic,and extends to medium density gradually;Two regimes in the headway-velocity relation are observed:At the constrained regime,the correlation between headway and velocity is weaker as the view becomes worse;At the free regime,pedestrians’ velocities under different visibilities conform to Gaussian distribution(VLT=0.3%～N(1.31,0.072),VLT=0.1%～N(0.71,0.092)and VLT=0.0%～N(0.45,0.102));Three regions in the density-flow relationship are obtained under different view-limited conditions:Free flow region,capacity region and congested region.The maximum specific flow under light transmittance of 0.3%,0.1%and 0.0%is 1.3s-1,1.1s-1 and 0.9s-1 respectively.To study the evacuation process under view-limited condition,we carried out two evacuation experiments under partially limited and completely limited view conditions respectively.From the experiment with partial view in a supermarket,some typical behaviors such as following behavior,helping behavior and finding the dependent during evacuation were observed from the experimental video.Afterwards we made questionnaire survey to further investigate the effect of partially limited condition on evacuation,and some discrepancy was found between survey result and experimental observation.Finally a multi-grid model considering these typical behaviors was proposed to simulate pedestrian evacuation under limited visibility.Simulation results of evacuation times in different scenarios agreed well with the experiment results,which validates the reliability of our model.In the evacuation experiment with completely limited view,the typical movement process is as follow:Firstly the participant moves to his or her face direction to look for the wall,and then when the participant touches the wall,he or she needs to select movement direction.In here the preference of choosing left-hand side direction is found in our experiment,afterwards the participant continues to move along the wall until the exit is found.It should be noted that different strategies of conflict resolution are adopted by pedestrians if conflict occurs during movement.Besides,T test result shows that the velocity of pedestrian moving along the wall is larger than that before finding the wall.In the next we built a multi-grid model for evacuation without visibility based on experimental observation and result analysis,and the model reproduced typical movement features of pedestrians under zero visibility.Finally we compared the evacuation results through simulation and found similarities and discrepancies between good and zero visibility condition.The similarities include:The initial distributions of pedestrians(random,uniform and centralized distribution)have little effect on overall evacuation time;Evacuation time increases with the increment of initial density in the room;Increasing the number of exits can decrease evacuation time,especially for zero visibility condition;The distributions of time interval in both situations show power-law relations.The differences are as follow:Due to larger movement velocity and shorter movement distance,evacuation under good visibility is much faster than that without visibility apparently;Increasing exit width can reduce evacuation time within limits for evacuation under good visibility while it has little influence on evacuation under zero visibility;High density exists in exit zone for evacuation under good view condition while it exists around the wall for evacuation without visibility.Through the above work we have a deeper understanding on pedestrian movement and evacuation under view-limited condition.In the following we explored and discussed how to guide pedestrians under limited visibility.Based on multi-grid model,an evacuation model with guiders was firstly established.Two types of pedestrians including guider and herding occupant were set in the model.The effects of guider number,guider type,guider distribution,guider velocity and guidance strategy on evacuation time were studied and the simulation results turned out:identified guiders are distinguished easily by other individuals,which is more beneficial to evacuation.In this case the evacuation time decreases by 7%when compared to that with unidentified guiders;The guidance range is expanded during the movement of guiders and the evacuation time under dynamic guidance reduces by 20%in contrast with that under static guidance.The optimal guider number exists in view of the human cost and evacuation time;The uniform distribution of guiders covers more area in the room and makes evacuation efficient,however the result is related to the distribution of herding occupants;Evacuation guidance is more effective when the speed of guider is about 75%of herding people’s speed in our simulation scenario;The performance of guidance strategy considering the distance to target pedestrian and number of people around the target is optimal when compared to other strategies;The communication and cooperation between guiders can avoid the conflict of guiding the same target,and the evacuation efficency increases by 6%.