The Inference Of Trip Purpose Based On Transportation Smart Card Data

Under the concept of “People-Oriented” in the urban public transportation system,it is necessary for the urban public transportation system to in-depth study of individual behavior.Transportation smart card data provides a wealth of basic data for the study of individual trip behavior,but the lack of trip purpose information prevents a deeper understanding of individual trip behavior.With the gradual diversification of residents’ trip needs,how to cost-effectively supplement the trip purpose in the transportation smart card data has become a research hotspot.This article uses transportation smart card data to extract individual multi-day trip activity information,and uses multi-day trip regularity characteristics as the main inference factor to construct a trip purpose inference model.Based on the distribution characteristics of trip activities at stations with different land functions,this paper discusses the possibility of irregular trip purposes under complex land use.The specific research content of this paper are as follows:(1)Shanghai transportation smart card data is selected as the basic research data,which is pre-processed for further study.The definition and type of trip purpose and the difference of different trip purpose and other related basic theories are introduced here.And the realization of trip purpose is elaborated.Rail transit passengers are selected as the research object to solve the problems caused by the limitation of transportation smart card data to infer the purpose of trip.The basic time and space characteristics of rail transit passenger flow are analyzed to provide a basis for simplifying the characteristics of individual trip rules.(2)Based on the pre-processed transportation smart card data set,the individual multi-day trip activity information is extracted.Firstly,the virtual transfer behavior that requires outbound transfer in the rail transit system is identified.Secondly,trip OD,stay time and stay activity mode are extracted to generate individual rail transit multi-day trip chains.Finally,based on the trip chain and the spatial relationship between the stations,clustering the trip stations to obtain the active area.The results show that converting individual trip station information into activity area information can reduce the error of multi-day trip law analysis.(3)The trip purpose inference model is proposed in this paper.Firstly,taking the activity area as the basic unit,the elements that affect the inference of trip purpose are extracted from two aspects: trip activity characteristics and multi-day trip regular features.Then,the distribution characteristics of the inferred elements are analyzed,and the trip purpose inference model is constructed heuristically.Data from the fifth Comprehensive Traffic Survey of Shanghai is used to verify the effectiveness of the model.The model proposed in this paper also identifies residents with part-time commuting purposes and residents who often use rail transit to commute one-way,which has certain significance and prediction for in-depth understanding of individual trip behavior.(4)According to the results of inference trip purpose,firstly this article analyzes the distribution characteristics of Shanghai rail transit loyal residents for various trip purposes.Then,using POI data,rail transit stations with independent land functions is filtered out.Combining with the characteristic distribution characteristics of trip activities arriving at different land function sites,the possibility of different irregular trip purposes under complex land use is discussed.