Study On Mechanisms And Simulations Of Pedestrian Flow Bottleneck

The influence of bottleneck on pedestrian movement is very important.Both flow-limited bottleneck and width-restricted bottleneck may affect the pedestrian flow.Using simulation software for pedestrian movement simulations is very popular in this field.However,the comparative study on the simulation results of different software is lacked until now.Therefore,based on the data of two large-scale pedestrian flow bottleneck experiments held in 2018 and 2019,this paper compares the simulation results of the same scenario made by three different software,and evaluates their advantages and disadvantages.Firstly,in order to investigate the mechanism of pedestrian flow bottlenecks,Pe Track is used to collect the microscopic and macroscopic data related to pedestrian flow characteristics.Firstly,the data of the flow-limited bottleneck experiment is analyzed.It is found that both the averaged value and the standard deviation of pedestrians’speeds grow with the increase of the distance from the bottleneck,and the propagation of stop-and-go waves could be observed.Secondly,the data of the width-restricted bottleneck is studied.The density waves and stop-and-go phenomenon could occur when the predetermined density is 4 ped/m~2,rather than the situation when it is 3 ped/m~2.And then,VISSIM,Any Logic and SUMO are used to simulate the influence of two different types of pedestrian flow bottlenecks.The microscopic and macroscopic data related to pedestrian flow characteristics are outputted.The results show that when the flow-limited bottleneck is considered,the statistics of pedestrians’speeds simulated by three software are different from that in the experimental results.The stable propagation of the stop-and-go waves could be reproduced by VISSIM and SUMO,while it is destroyed by the abnormal effects of the road corners in Any Logic.In the simulation of width-restricted bottleneck,some features simulated by Any Logic are close to that in the experimental results.When the predetermined density is 3 ped/m~2,the simulation results of SUMO are also close to the experimental results.But they cannot be realized when the predetermined density is 4 ped/m~2.On the contrary,the simulation results of VISSIM are quite different from that in the experiments.In a word,the three software have different advantages and disadvantages when simulating the pedestrian flow bottlenecks.Due to the limitation of the software functions and the deficiencies of the microscopic pedestrian flow models embedded,none of them is suitable for high-density pedestrian flow simulations,and the experimental results cannot be perfectly reproduced.All the above conclusions could provide some references for future work of pedestrian flow simulations.