Research On Correction Method For City Expressway Traffic Flow Multi-source Data

The contradiction between the supply of urban road infrastructure and traffic demand is increasingly prominent. The problem of traffic congestion is growing. Currently, relying solely on widening roads,increasing the density of the road network and other infrastructures have been difficult to solve urban traffic congestion problems. However, research and development of intelligent transportation system is a feasible way. Advanced traffic management system is an important part of intelligent transportation systems.Thereinto,the identification of quality and correction of traffic flow data is the most basic and critical component of the advanced traffic management system. Then how to improve the accuracy and the instantaneity of urban roads traffic flow data becomes extremely important.This paper puts forward the combination of single-source data correction with multi-source data correction to improve the accuracy of the road traffic flow data. Firstly,the paper studies quality identification and correction of single-source data on different detection data. The paper is used to identify and classify abnormal data based on traffic flow theory and threshold theory, then applies historical trend method、time series method and the characteristic of the local stability on traffic flow to repair missing data,and propose improved Etkin interpolation algorithm to correct erroneous data. Secondly, the paper puts forward BP neural network based numerical optimization to correct multi-source data,then selects five numerical optimization methods,such as BFGS quasi-Newton method, quasi-Newton tangent method, Fletcher-Reeves conjugate gradientmethod,Polak-Ribiere conjugate gradient method and the Levenberg-Marquardt algorithm to improve BP neural network, and respectively performs comparative analysis on their relative error, run-time, iterations. This paper selects multi-source testing data of the second ring road in Beijing for example.