Estimating Urban Resident’ Public Transport Ods From Smart Card Data Using Bayes Inference

Vigorously promoting public transport and building the transit metropolis are the key measures to solve urban diseases such as traffic congestion,environmental degradation and resource shortage.To build a transit metropolis,it is necessary to understand the travel behavior and characteristics of urban residents.The traditional way to understand urban residents’ travel is mainly residents’ travel survey,which has disadvantages such as high cost,long cycle,and relative lag.At the same time,as static cross-sectional data,it is insufficient in analyzing the characteristics of residents’ travel.With the popularization of smart public transportation card swiping systems,the emergence of massive public transportation travel big data provides a fast and low-cost solution for understanding residents’ travel behavior.However,due to the limitations of card swiping behavior and smart card storage systems,the current bus swiping system in most cities does not collect and store the information of residents’ pick-up and drop-off stations,which hinders the understanding and analysis of residents’ bus travel behavior.Faced with the problem that public transportation big data cannot directly reflect the travel behavior of urban residents,the purpose of the study is to use smart card data to realize the OD inference of residents’ public transportation trips by comprehensive urban planning and transportation disciplines.The main idea is to comprehensively use bus smart card data,bus trajectory data,urban construction land information and social management grid data.Combining the basic situation of the data to determine the problems and needs that need to be solved to infer the residents’ public transport travel OD,and design the data processing algorithm for the problems,mainly including the following parts.Based on the matching and recognition of time and space information,the vehicle arrival timetable is obtained,and the vehicle arrival timetable and smart card data are further combined to extract the pick-up station of residents’ public transportation trips.Taking the residents’ two consecutive bus trips as the research object,the concept of potential path space in time geography is used to analyze the range of activities between consecutive bus trips.The scope of residents’ activities is understood as the social management grid unit covered by residents in an elliptical shape constrained by the restriction of the activity time around the two travel stations.Use Bayes inference to combine the distance attenuation effect and the built environment characteristics of different activity ranges to obtain the influence of the residents’ next bus trip on the probability of getting off the station for the previous trip,and then get the possibility of residents traveling between bus stations.Infer the distribution of the residents’ trip probability among the social management grid units from the residents’ bus travel behaviors between stations,and extract the residents’ travel OD with the social management grid as the basic unit.The basic idea of using subway data to verify the effectiveness of the method based on Bayesian inference and extraction of residents’ drop-off stations is proposed using subway data at the individual level,and the above method is verified by using Shenzhen subway data,and the accuracy of the method of inferring drop-off stations is about 90%.Based on the above method,the data of the bus smart card in Shenzhen in October 2017 is processed to realize the extraction of the OD of residents’ pick-up stations,drop-off stations,and social management grid.Collect travel behavior at the community level to analyze group travel characteristics.The inference results can directly reflect the strength of the interaction relationship between public transportation in Shenzhen’s communities,and reflect the differences and changes in the relationship of public transportation between communities in different regions and at different times.It is helpful to understand the status quo of urban residents’ public transportation needs and explore residents’ travel characteristics.