Research On Passenger Flow Analysis Algorithms Of Urban Rail Network Based On AFC Data

With the expansion of urban traffic scale,the urban rail transit network system has gradually improved.Under the situation of "one ticket transfer",it is the basis of passenger flow prediction and operation management to grasp the dynamic distribution law of passengers and monitor the change of passenger flow in road network.The AFC data provided by the Automatic Fare Collection system contains information such as the OD pair of each passengera well as the arrival and exit time,it is possible to mine individual travel information from AFC data and grasp the distribution law of passengers,so we can mine the potential information of these data and analyze the passenger flow distribution characteristics of the road network.At present,most of the researches on passenger flow analysis of urban rail transit focus on multi-path selection and passenger flow assignment to match the specific travel paths of passengers,but no further track the trajectory of passengers and observe the distribution of passenger flow in real time..In this context,based on passenger flow matching,this study amis to analysis the spatio-temporal characteristics of passenger flow distribution in road network,firstly,determining the route of each passenger and analyzing the mode of operation of the passenger according to AFC data and train operation data,and data mining,Then,according to the time axis,the passengers are matched to the specific space position of the road network,and the traffic volume of the stations and sections in the road network is analyzed and studied.Finally,providing the implementation scheme of dynamic demonstration system,this system realization of dynamic demonstration of passenger flow distribution and congestion state in urban rail transit network.Based on the above assumptions,this paper proposes a path matching algorithm based on FCM clustering,fuzzy matching and density probability proportional correction,and a passenger flow analysis algorithm for urban rail network based on passenger flow matching.For the problem of missing cluster center and unable to match one by one in path matching algorithms based on AFC data,this paper takes the effective path as the classification center,and divides the effective path into two kinds of experimental sets,which can be divided into two types according the benchmark travel time similar,the fuzzy matching algorithm is innovatively introduced to the non-separable effective path to measure the affinity density between the sample set and the classification center by the fuzzy function and realize the one-to-one matching between passengers and paths.In order to improve the precision of reference travel time estimation,the fitting algorithm of walking time in and out station is modified,and the fitting value of all stations is replaced by the average value of some stations,the density probability proportional correction function is used to effectively reduce the experimental error and improve the precision.The experimental results show that the proposed algorithm can ensure the precision and realize the one-to-one matching between passengers and paths at the same time.For the problem of passenger flow source diversity in road network site and cross-section,manual survey time-consuming and unilateral statistical inadequacy,this paper proposes a passenger flow Analysis algorithm based on passenger flow matching.This algorithm determines the passenger’s travel path based on the result of path matching,on the basis of this,the AFC data,operation data and the estimated value of probability distribution parameters of effective path benchmark travel time are used to mark and assign values to road network stations and sections and match passenger flow on the corresponding sites and sections on the travel path by time,analysing and researching on passenger flow of Statistical stations and sections,then,the passenger volume of the station and section is analyzed and studied,and the research is applied to the dynamic demonstration of the distribution and congestion of the road network passenger flow in the Beijing urban rail transit line map.The experimental results show that the algorithm can well analyze the distribution characteristics of road network passenger flow,and it is convenient to observe the change of road network passenger flow,which has the practical significance.