Research On The Calculation Of OD Path Selection Probability Based On Taxi GPS Data

Drivers of urban vehicles often need to choose routes based on OD(Original-Destination),where the shortest route is usually an important reference for selection.However,in an urban environment,the shortest path does not mean the minimum travel time.This is because the traffic state of the OD path has spatiotemporal characteristics,that is,the traffic state is closely related to the change of time and the flow of the road.The vehicle always senses the OD route traffic state from the moment of departure to select the route,but the route traffic state may change after departure.Urban taxi drivers are the type of people who are most sensitive to changes in traffic conditions with spatiotemporal characteristics.It is generally considered that the experienced drivers among them have the best route choices under specific spatiotemporal conditions.Using GPS data to study the OD path selection of these people can serve the OD path selection of urban driving travelers under a specific time and space.The research content and results are as follows.(1)Realize the construction of the urban road network and the road matching algorithm based on the grid map of the taxi passenger time series GPS trajectory.The digital road network structure is constructed based on the OpenStreetMap map data in Xi’an;the taxi GPS trajectory data is divided into passenger OD time series trajectory sets,and they are matched to the digital road network to obtain a taxi passenger path set.(2)A method for extracting the threshold of unobstructed passenger speed is proposed.The driver’s income is used as the basis for judging the taxi driver’s experience,and the K-Means clustering is used to divide it into three levels: low-income,middle-income and high-income.Logistic regression was used to analyze the relationship between passenger speed and income level,and then determine the speed threshold for dividing the income level of low-income drivers and middle-high income drivers.The speed threshold can evaluate whether a specific set of OD paths is a smooth judgment standard.(3)A probability calculation method for path selection for a given departure time and specified OD is proposed.According to the time zone where the departure time is,the unobstructed route set is filtered and matched with the specified OD to generate an OD alternative route set,and then the driver’s selection probability for each reachable route in aspecific time period is obtained according to the alternative route set.The mining of taxi GPS trajectory data can obtain the OD path selection probability of experienced middle-and high-income taxi drivers in a specific time and space.The greater the probability of selection means that the road network environment in a specific time and space and changing traffic conditions The greater the probability that the OD route has a shorter transit time,which has a guiding significance for the OD route selection behavior of urban driving travelers.