Traffic State Identification Modeling And Prediction Of Traffic Status In Loop Network

At present,the formation of the urban business circle network is the result of natural selection of the development of the urban central area.The network loop network is composed of two basic modules: the single circular road in the center position,and the radial distribution road of the connecting ring road and the external road network.The two modules assume different traffic functions and interact with each other,forming the core of the business circle network.Due to the geographic speciality and special features of the loop network,the loop network has extremely high importance and vulnerability in the road network system of the entire region.In order to grasp the traffic status of the loop network and avoid the instability of the loop network traffic,this paper studies the data preprocessing,traffic status identification and traffic parameter prediction.Because the control and early warning of ITS requires the input of real-time traffic parameters,due to the inherent defects of the detector and the error interference during the detection,the detection of the traffic operation parameters is distorted to some extent.Based on the integration of existing traffic operation parameter pre-processing methods,this paper constructs a set of screening and repair methods for traffic parameter anomaly data.The data after screening and repair have higher credibility than the original data.The identification and modeling of traffic operation state is based on the analysis of traffic operation characteristics.Due to the structural differences between the ring road network and other networks,there are also some deviations in the characteristics of the road traffic.On the basis of analyzing the influence range and structure characteristic of the loop network,the traffic state is divided into three states of free flow,synchronous flow and clogging flow based on the detection data and three phase traffic theory,and the corresponding system of microscopic traffic state and macro traffic state is established,and the macroscopic traffic state is classified and selected.Class index threshold.Finally,on the basis of wavelet analysis,an optimized BP neural network is constructed based on the traffic flow theory model.The traffic parameters are predicted in real time using the historical data,and the traffic parameters are combined with the model of traffic running state inference.The purpose of advancing the state of traffic in advance.