Passenger Flow Calculation For Bus Stations Based On IC Card Data

Public transportation passenger flow in bus station is the basic data of urban transit planning, operation scheduling and organization. However, both the expansion of city size, and the rise and frequent changes of bus routes and stations have brought many difficulties to investigate public transportation using traditional manual method. And transfer passenger flow statistics in transit site is related to the traveler’s personal choice behavior, but the questionnaire or inquiry survey consume much and the data quality usually can’t meet the demand. With public transportation IC card widely used in many major cities in China, data processing and analyzing based on IC card have widespread operation platform and a huge data support, which make it possible for passengers flow statistics using information technology.Currently, the main goal of IC card data analysis is to obtain the bus passenger information for bus operators in decision-making and residents travel features for transit planning. But the bus IC card was originally designed for the ticket settlement, so IC data of segmented pricing lines is missing the boarding time, and integrated ticketing lines is missing the boarding and alighting station and time respectively. Moreover, to provide convenience for passengers,53%of bus lines credit IC card only once after getting on in Beijing, which lead to the rule of passenger flow in bus station hard to gain by IC card data directly.Taking Beijing bus IC card data as the foundation, this paper conducts research on acquiring public transportation passenger flow in bus station, solving problems like incomplete travel information and the IC card data record data error, the main contents include:Firstly, brief introductions to the bus IC data component and characteristics are given, including three types of IC card data: segmented pricing, integrated ticketing and subway, and data about subway is the most comprehensive. Propose the problems and corresponding solutions of IC card data including field loss, data errors and zero station, for field loss to take padded measures, for indentified error data taking it as field loss, for zero station passenger flow redistributing and forming the actual passenger flow in each station and storage them, so as to improving the quality of IC card data. Then, the data analysis system architecture and public transit IC card are established targeting at calculating the passenger flow volume in bus station. Secondly, because of the lack of the boarding and alighting station data, this paper distinguishes bus lines and classify the single trip IC card data by the shortest distance system clustering method. In each category, match the first time the IC card is used and bus arrival time of the GPS data to determine the boarding stations. In the case of no GPS data, at least two clustering are identified by the transfer information, and clustering is determined by statistical pattern recognition method. For the complete bus trip, determine the last getting off station according to the following boarding station, for the incidental non-bus travel, this paper proposed the concept of attracting strength of station, to compute the probability of the station.Thirdly, put forward distance restrain and time restrain of transfer identification, namely, two records of swiping the card between different bus routes and different ticket systems, if the transfer walking distance is within1000m, and the time interval is within the range of a variable parameter which consists of riding time, walking time for transfer and waiting time in transfer, which would be considered as only one transfer, using trip mark field to present the results of identification. The two times of swiping the same IC card with the same trip mark is considered as a transfer, otherwise it is considered as a second trip. Then this paper verified all kinds of methods through survey data, and then compare it with transfer identification method of constant parameters.Fourthly, propose passenger flow projection method in bus station based on the results of transfer indentification, including arrivaling flow, transfer flow, and leaving flow. Then acquire transfer flow matrix between the bus lines through accumulating transfer flow in each station, and calculate the bus transfer coefficient, rate and average waiting time to evaluate the level of public transportation transfer service.The results show that method of passenger flow volume projection in bus station based on IC card data in this paper, which can effectively deduce bus passengers boarding and alighting station, and indentify transfer behavior to get bus stop flow volume and provide decision support for transportation planning and scheduling.