Research On The Method Of Judging Pedestrian Traffic State In Urban Rail Transit Passage

Passages are important facilities for pedestrians in and out of urban rail transit stations.During holidays or peak periods,a large number of pedestrians are often stranded in rail transit station passages.Some accidents are likely to occur,which increases the potential safety hazards for pedestrians.Research on the classification and discrimination of pedestrian traffic conditions in urban rail transit channels is helpful for rail transit operations and management departments to obtain information on passenger flow operating conditions in the corridors,and decision makers can make scientific evaluations of passenger flow operating conditions for timely implementation A series of measures such as passenger flow evacuation guidance and emergency information release,so as to avoid the occurrence of pedestrian traffic safety accidents in the passage,so that passengers can travel on rail transit more comfortably and comfortably,thereby improving the level of rail transit service.The main research contents are as follows:(1)Define the pedestrian traffic status in the rail transit channel,analyze the relationship between the pedestrian flow parameters,select density,speed,and flow as the distinguishing parameters of pedestrian traffic status,and select pedestrians at two rail stations in Chongqing The channel performs video shooting and collects the required pedestrian traffic flow parameters.(2)Select the relative division standard to classify the pedestrian traffic status,introduce the principle of the fuzzy C-means(FCM)clustering algorithm,analyze the parameters of the FCM algorithm,and determine the pedestrians according to the clustering validity function of the FCM algorithm The number of divisions of the traffic state;for the FCM algorithm due to the random allocation of the initial clustering center,the clustering results are unstable and easy to fall into the local optimal problem,the genetic algorithm is introduced to optimize the selection of the initial clustering center for the FCM algorithm,and the pedestrian traffic is increased.The reliability of state clustering analysis,a pedestrian traffic state discrimination model based on random forest and a pedestrian traffic state discrimination model based on BP neural network are constructed.(3)Based on the pedestrian flow data obtained from the survey of the pedestrian passages at the two stations of Chongqing Nanping Railway Station and Hongqihegou Railway Station,an example analysis of the pedestrian traffic status identification method in the urban rail transit passage established in this paper is carried out,and the results show that,The fuzzy C-means clustering algorithm based on genetic algorithm has a fast convergence speed and better stability;using precision,recall,and running time as evaluation indicators,the RF algorithm and BP algorithm are run many times,and the results show that the random forest algorithm and The average precision and recall of the two algorithms of BP neural network after multiple runs can reach more than 93%,and the average running time of the two algorithms is very short,which shows that the random forest and the BP neural network algorithm It is feasible and reasonable to be used for the judgment of pedestrian traffic status.