Extraction Of Passenger Traffic Congestion Based On Subway Card Data

With the rapid pace of urbanization,the city congestion problem becomes more and more serious.Development of public transit has been an effective measure to resolve the heavy traffic jam in many big cities all over the world.Compared with other public transportation,urban subway has the advantages of high speed,economy and vast capacity.Large numbers of people take the subway to work,which causes the passenger congestion in subway train or at the station in the morning and evening peak hours.And the potential safety hazard is growing.In case of emergency like stampede,explosions or terrorist attack,the consequences will be severe.Therefore,it is grasping subway traffic in advance that is crucial to public security on the underground.At present,research about the number of sub way passenger focuses on passenger flow forecast,passenger simulation and real-time passenger flow.But the study scale is for a certain station in a certain period of time not for a certain subway train.In fact,passengers produce millions of subway smart card data,that is,when passengers tap in and out subway system,the time and station name will be recorded.Subway smart card data is a kind of spatiotemporal data.For this reason,we propose a scheme to calculate the number of passengers in the train and waiting at the station platform based on subway smart card data.This paper consists three parts:data processing,algorithm introduction and case study.Taking Shanghai as an example,the topological vector map of Shanghai subway station and line is constructed.In addition,the trip data model is developed,through which the chaotic raw data could be coded to trip data.As we all know,there are two directions in the subway line.Then,in order to judge passenger's trip direction the trip data would be divided into two categories:the passenger transfers from one line to other line or he does not.When the passenger does not transfer,we could judge his trip direction based on the site where he gets on and the site where he gets off.However,if the passenger changes line,all his station that he passes through would be inferred utilizing the K shortest path algorithm.Time clustering algorithm is used to find out the earliest time when passengers tap out the station continuously from the same train.Furthermore,a schedule data mining algorithm is designed to get the timetable for each train using the handled trip data.Finally,Shanghai Metro Line 5 and Line 8 are illustrated as case study.The experiment results include timetable and the passenger flow for each subway train.What's more,the passenger flow would detail to the number of waiting passengers at the station platform,the number of passengers who get on the subway train,the number of passengers who get off the train and the number of passengers in the train when the train leaves,which not only can be used in the research of the characteristics of rail transit travel,but also provide support to improve the level of public transport services.Besides,density of standing passenger in the subway train is defined to describe the crowding in the train.Also,we collect the number of passengers in field and evaluate the experiment result to ensure the reliability of the program.