Urban Link Travel Time Estimation Using Vehicle OD Data

It is great significance for the urban public transport management and gathering events to estimate urban link travel time in the process of urbanization.In recent years,with the rapid development of wireless sensor and mobile positioning technology,mobile communication and mobile computing technology has been widely used in various fields.City mobile objects in the process of free movement with the network to maintain interconnection,and exchange data with the server,resulting in a large amount of location information with the spatiotemporal label.Therefore,the implementation of urban link travel time estimation has a large-scale data base.Aiming at the time and space data carried out the research based on the existing methods of estimating the passage time of the road network and the details of the road section.1)Construction of framework model which based on the OD data to estimate the urban link travel time.Based on the existing urban link travel time estimation method,the frame model of urban link travel time estimation based on vehicle OD data is summarized.It mainly includes three steps: map matching,reasonable path establishment and urban link travel time estimation.2)Based on the actual situation of the multi-lane in each section of the road network and the interaction between the left lane and the right-hand lane of the same section,the paper puts forward the method to estimate the transit time of the road network section of the determined two-lane model.The relationship between the two-lane model of the polynomials is analyzed,and the optimal two-lane transit time relationship is determined by comparing the influence of different orders on the results.Based on the established urban link travel time estimation,the paper proposed the methods of the urban link travel time based on the established two-lane model.3)In order to solve the shortcomings of deterministic two-lane model and improve the fine degree of road network,a random variable is introduced to describe the dynamic change of road network.In this paper,a random variable model is proposed to estimate the passage time of road network with the hypothesis that the travel time on each section and the traffic light waiting time for each intersection are random variables and subject to a certain normal distribution.According to the relevant theory of Bayesian learning,the factors and the probability that influence the passage time of the road network segment are analyzed,and the passage time of the road network is fitted through the parameter estimation method with hidden variables.