Research On Methods Of Extracting Commuting Trip Characteristic Based On Public Transportation Multi-source Data

Commuting travelers are the critical subject of public transportation services,while accurately grasp the trip characteristics of them is a basis of several publictransport measures implementation, such as system construction, network evaluationand optimization, travel demand identification, and priority policies development. Inaddition, with the application of advanced data acquisition technology in APTS(Advanced Public Transport System), different types of massive public transportoperation data are accumulated. These data could record the run trajectories ofvehicles and describe the travel courses of public transport users in temporal andspatial, which provide a good foundation for extracting the characteristics of publictransport trip. However, how to extract the commuting trip characteristics frommulti-source massive data by matched and associated effectually had not formed a setof complete method system.Based on the analysis of public transit smart card data, real-time bus GPS data,bus routes data and railway station statics data, with the intention of represent thetravelers’ public transportation path, the study proposed several methods ofmulti-source data matching and associating to make a data prepare for tripcharacteristics extraction.From the comprehensive reviews of trip chain concepts, the study put forwardthe definition of "public transportation trip chain (PTTC)" by introducing the conceptof journey stages, which realized a detailed expression for the public transportationcourse on the basis of traditional travel chain. Then, with comprehending theingredients of PTTC and taking the multi-source data preparation as a foundation, thestudy proposed a “Four Phases” method of extracting commuting PTTC. They wereextracting the PTTC’s structures, identify the commuting trip, ascertain the origin anddestination of journey stages, as well as calculate the stages’ travel time and distance.For each extraction phase, the study constructed a detailed processing rules andprocedures, which involved finding the discriminated threshold of different transfertypes for connection journey stages, marking the commuting trip in smart card dataaccording to the similar structures of PTTC, and deducing the bus passengers’boarding and alighting station by made use of bus arrival time, spatial relationship ofmulti-modal sites and features of commuter trip rules. Accordingly, it would berealized the complete extraction of ingredients of commuting PTTC. Furthermore, afield survey was designed to verify the precision of the extraction model, and theverification results indicated that the model showed a good accuracy for publictransport trip characteristics extracting.Finally, by taking advantage of Beijing public transportation multi-source data of five consecutive workdays, the study completed the extraction of commuting tripcharacteristics for a specified day. Meanwhile, the distribution of journey stages,structure features of commuting PTTC, travel distance and time of commuting trip,and indicators of transfer characteristics were analyzed according to the extractionresults. The case study indicated that the extraction model of commuting PTTCpossessed a nice applicability in practice, which would provide targeted decision datasupport for evaluating bus operation and improving public transport service level.