Research On Prediction Of Freeway Traffic Congestion Based On Ground Induction Coil Data

The prediction of freeway traffic congestion is a key supporting technology freeway intelligent management.It is of great significance to reduce the occurrence of traffic congestion and give full play to the characteristics of freeway safety,speed and efficiency.To realize the accurate prediction of the expressway traffic state,firstly,it is necessary to obtain the predicted value of the traffic state characterization parameter of the freeway section in the future;Secondly,use the predicted value of the traffic state characterization parameter of the freeway section to judge the road traffic state for a period of time in the future,so as to realize the prediction of the traffic congestion state.Firstly,a short-term prediction model of the characteristic parameters of the traffic congestion state is established.The short-term prediction of freeway traffic state characterization parameters can provide necessary data support for the prediction of traffic congestion state.The reasonable selection of characterization parameters and the accuracy of prediction are of great significance to the judgment of traffic state.In order to overcome the failure of the existing short-term traffic parameter prediction methods to fully consider the randomness and nonlinear characteristics of traffic flow,a short-term traffic flow prediction model optimized based on ensemble empirical mode decomposition algorithm(EEMD)combined with wavelet analysis is proposed.Firstly,the EEMD algorithm is used to decompose traffic flow data into multiple intrinsic mode functions(IMF)and a residual component(RES);Secondly,through the autocorrelation analysis of each component,by using the autocorrelation function characteristics of the noisy signal to filter out the noisy components in each component,and further use the wavelet analysis algorithm to process the noisy components;Finally,two types of model construction methods are proposed:(1)Reconstruct the IMF and Res after wavelet threshold processing,and input them into the short-term traffic flow prediction model,and the model output is the final prediction value;(2)Input the IMF and Res processed by the wavelet threshold into the short-term traffic prediction model,and the output of the model is the predicted value of each component,which will be the final predicted value after reconstruction.Then,a traffic congestion state discriminant prediction model is established.The accurate discrimination of traffic congestion can provide an important basis for highway traffic management departments to formulate scientific traffic management plan,and it is also of great significance.According to the traffic parameter data obtained from the traffic parameter prediction,the freeway traffic congestion state prediction model was constructed.Firstly,the entropy method is used to weight the selected traffic state discrimination parameters to realize the quantitative description of the importance of the state discrimination parameters,that is,by calculating the entropy value of each type of traffic state discrimination parameter,the weight of each type of discrimination parameter is obtained.Secondly,the Euclidean distance metric function is improved by parameter weighting to improve the fuzzy C-means clustering(FCM)algorithm,and the improved FCM algorithm is used to cluster analysis of the traffic state characterization parameters,and divide the data set into four traffic state sample data sets{unblocked,stable,congested,blocked};Finally,in view of the shortcomings of the existing traffic state discrimination methods in terms of the degree of refinement,rationality and stability of the discriminated states,a traffic state discrimination prediction model based on support vector machine multi-classifiers is constructed.Finally,based on the traffic flow parameters of the Whitemud Drive freeway in Canada,the short-term traffic flow parameter prediction model established in this paper is experimentally analyzed.The results show that the two types of short-term traffic flow parameter combination prediction models constructed in this paper have good prediction accuracy and practicability.Furthermore,based on the predicted characterization parameters of traffic flow,an experimental analysis of the highway traffic state discrimination model constructed in this paper is carried out.The results show that the improved fuzzy C-means clustering algorithm based on entropy method has good stability and high classification accuracy;the traffic state decision model based on multi-class support vector can ensure the accuracy of classification.