Bayesian OD matrix estimation method with multiple data types

Abstract

The present invention discloses a Bayesian OD matrix estimation method with multiple data types. The method comprises the following steps of: the step 1, obtaining four data types which are a road section flow, a flow of part of paths, a turning flow of an intersection and a road section vehicle travelling speed, performing layering and screening of the four data types, and obtaining relatively independent traffic data on a network; the step 2, respectively calculating likelihood functions corresponding to the road section flow, the flow of part of paths, the turning flow of the intersection and the road section vehicle travelling speed according to the relatively independent traffic data on the network obtained in the step S1, and obtaining a posterior distribution form of Bayesian OD matrix estimation; and the step S3, designing a Markov Chain Monte Carlo sampling method, and solving the posterior distribution form of the Bayesian OD matrix estimation in the step S2. The Bayesian ODmatrix estimation method with multiple data types is helpful for improvement of accuracy and a technology application range of the current OD estimation.

Description

Technical field: [0001] The invention relates to a multi-data type Bayesian OD matrix estimation method, which belongs to the technical field of urban traffic network analysis and optimization. Background technique: [0002] The OD matrix describes the traffic volume from all origins to all destinations on the traffic network within a certain period of time, reflecting the traffic demand of travelers; OD traffic is the basic input data for urban traffic planning, management control, etc. An important basis for scientific research work. The traditional OD matrix acquisition is time-consuming and laborious, and the accuracy is not high. The OD matrix can be estimated or reversed by observing the traffic information on the road network. The high-precision OD matrix estimation results can lay a good foundation for other related processing work. [0003] However, it is often impossible to uniquely determine the OD matrix based solely on the observed road traffic flow, especially...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a Bayesian OD matrix estimation method of multiple data types, to solve the OD matrix estimation problem in the urban traffic network analysis, by obtaining relatively independent traffic data on the network to obtain the likelihood information of variables, Then use the Bayesian formula to infer the posterior distribution of all variables (including OD matrix elements and model parameters), and finally achieve the goal of improving the accuracy and effectiveness of the basic data of the urban traffic OD matrix

Images