Passenger Flow Estimation Based on Smart Card Data in Public Transit

Abstract

The high-efficiency operation of an urban public transport system is increasingly based on advanced management and techniques. With the development and application of intelligent transportation technology, more information is available to and used by urban bus operators. The smart card has become citizens' most popular transaction mode on public transportation systems. It provides a source of a massive amount of continuous and dynamic information regarding the travel behavior of public transit users. Most smart card data is stored in the boarding information when the passengers board, but no alighting information is stored. Boarding stops can be estimated by matching the transaction data and global positioning system (GPS) data of the bus. In order |to make good use of the smart card data and analyze citizens' bus travel behavior, the authors propose an Alighting Stop Inference Model. Based on the assumption of urban bus travel chain closure, by considering the Euclidean distances between the next transit's boarding stop and the possible alighting stops of the current transaction trip, the most likely alighting stop would be given if the model's condition is satisfied. On this basis, the authors propose a method of determining the transfer behavior in bus rapid transit (BRT) channel and estimating the alighting stop with transfer behavior in BRT. Thus, the authors obtain the origin and destination (OD) information of passenger flow among public transit lines of Guangzhou City.