Urban Rail Transit Passenger Spatio-tempral Trajectory Inference Considering Train Load

Urban rail transit,the operational network of which has witnessed much improvement,has gradually become one of the most important transport modes for residents to commute,shop,travel,etc.However,due to the fact that AFC(Automated Fare Collection)data cannot record the travel trajectory of the passengers,the traditional passenger flow assignment methods were found to be insufficient to estimate the sectional passenger flow at the train level.And the increase in network complexity makes it even more difficult to grasp and analyze the spatio-temporal distribution of congested passenger flow during peak periods,which brought huge challenges to the safe operation of the subway.In this context,this paper proposes an urban rail transit passenger trajectory inference method that considers train load.Taking AFC data and train schedule data as input,designing an efficient path search algorithm under the bestfirst strategy,constructing a probability model for estimating the passenger feasible itineraries based on travel time elements distribution,and innovatively adding a penalty function for train passenger load based on the probability update mechanism.The research provided support for refined passenger flow analysis.The detailed contents and results of this dissertation are as follows:(1)Explain the research background and significance of passenger trajectory estimation,demonstrate the necessity of assigning passenger trajectories to specific trains,and summarize the current research status in related fields and the characteristics of different passenger trajectory estimation methods.(2)From the perspective of time and space,analyze the characteristics of passengers’ travel behavior at the three levels of network,line,and station,and summarize the distribution of morning and evening peak passenger flow.Take Chengdu’s rail transit network as an example to draw an OD distribution map,Station location map,roughly calculate the bottleneck section of the road network,analyze the rules of passenger flow,and provide a basis for the following article.(3)Introduce the existing efficient path search algorithm,take the urban rail transit network operation as the background,propose the definition and basic assumptions of the effective path,and propose the road network reconstruction rules to simplify the urban rail transit network into a directed weighted network and construct the efficient path search algorithm of urban rail transit under best-first strategy,and an inspection algorithm based on EM clustering is designed to validate the results.(4)Describe the passenger spatio-temporal trajectory estimation problem in detail,and put forward several basic assumptions;based on the definition of time nodes in the passenger travel process,the travel time is decomposed,and the concepts and influencing factors of each time element are proposed;construct probability density functional of each time element,clarify the parameters to be estimated of each model;based on the decomposition of travel time and the penalty of train passenger load,construct the probability model of passengers’ feasible travel plan selection;design the itinerary search method,the time element distribution parameter estimation method and probability update mechanism.(5)Taking Chengdu’s rail transit network and transaction data as an example,search and generate the effective path set of the entire network OD pair,and verify the two typical ODs;based on the passenger trajectory estimation results of a single ride plan,analyze the station performance Outbound time distribution,inbound time distribution and transfer time distribution;take three typical passengers as examples to verify the selection probability model of the ride plan based on the train passenger update;carry out the trajectory estimation of the passengers on the whole network on the day,and analyze the early statistics.Temporal and spatial distribution of peak passenger flow and cross-section passenger flow.In this paper,by studying the passenger travel behavior of urban rail transit,an efficient path search algorithm under the best-first strategy is proposed,and a spatiotemporal trajectory estimation method based on train passenger load penalty is constructed.It can analyze the passenger flow in urban rail transit in the level of trains.The temporal and spatial distribution in the network provides a basis for the refined passenger flow analysis and operation management of operating units,which has certain practical significance.