Research On Repair Method Of Highway Traffic Flow Fault Data Based On Spatio Temporal Relationship

With the convenient traffic conditions traffic volume has been increased rapidly, thus inducing city size and road network scale gradually expanding. Coupled with the advent of "metropolitan area", it makes the connection between Citys or between metropolitan area and other regions pretty closely. Meanwhile, highway has become the most important carrier to ease commutes.Now it is an urgent problem that how to gain useful information from plentiful traffic flow data sets collected by the highway monitoring equipment, and it has become the basis of accurate traffic conditions prediction, the improvement of traffic control measures and induction the of high-speed traffic.This paper will summarize the research status of traffic flow failure data repair methods, in order to improve the precision of data restoration model on the basis of comprehensive consideration of traffic flow data in space and time characteristics. Thus it can provide accurate and effective basic data for the subsequent module of traffic prediction and induction control. Therefore, this article analyses the characteristics of traffic flow data qualitatively and quantitatively from three aspects of the time, space and stability characteristics, providing a complete and effective data support for the repair model of traffic flow data; Considering the propagation characteristics of traffic flow in lengthway space, this paper proposes a method of space correlation estimation of traffic flow that under the influence of the highway traffic state on the basis of stationary and non-stationary time division; in view of the method above the spatial correlation estimation index is determined, which is used as an input parameter of fault data repair model; through the analysis of research status, this article intends to apply a spatio-temporal interpolation method based on 3D shape function to traffic flow data recovery field, with building spatio-temporal datasets, unifying space-time dimension of traffic flow data. The spatio-temporal interpolation method based on 3D shape function is improved in basis on the sequential smooth processing,which is used to repair traffic flow fault data. Finally,the spatio-temporal internolation based on stationarity is compared with the spatial interpolation method and time series analysis.The major works of this paper are listed as follows: on account of the space-time characteristics of traffic flow data on basic freeway section, a method of space correlation estimation of traffic flow that under the influence of the highway traffic state is proposed, and on this basis, a spatio-temporal interpolation model based on 3D shape function is built to repair traffic flow fault data. Through example validating, the proposed method can effectively repair the fault data, thus providing more effective and comprehensive basic data for traffic forecasting and inducement control.. Key words:...