Research On Urban Road Network Key Sections Identification And Short-term Traffic Status Prediction

It is very important for the operation and management of a city to monitor the running state of the urban road by using intelligent transportation system(its).It can help the urban planning and traffic management departments to better understand the traffic condition of the city by accurately identifying the key road sections in the urban road network,distinguishing the traffic state of the key road sections,and making short-term forecast of the traffic state,take timely measures to optimize the urban traffic system,improve urban traffic efficiency and safety.However,because of the limited data collection and the single index,the method of identifying the key road sections in the urban road network still needs to be improved.In addition,how to distinguish the traffic condition and improve the accuracy of state prediction is still worth studying.In order to solve the problems in the above research,this paper uses the license plate recognition data to study the method of urban road network critical section identification,traffic state identification and traffic state prediction.The main contents include:(1)mining the license plate recognition data to extract the traffic parameters in the urban road network.The pre-processing method of license plate recognition data and the extraction method of traffic parameter data based on vehicle track are studied.The traffic volume and the average travel speed are extracted,and the spatio-temporal characteristics of the road network in the study area are analyzed.The results show that the traffic flow in the study area has an obvious morning and evening peak,and the traffic flow correlation is strong.(2)in the research on the method of Critical Road section identification in urban road network,in order to solve the problem of single index and method of Critical Road section identification,in this paper,the PageRank-entropy weight-TOPSIS model is established from the angle of link physical characteristic index,link traffic demand index and road network Topology Index.The PageRank algorithm is used to mine the traffic transfer information in the road network,and the entropy weight-TOPSIS is used to construct the evaluation system of the importance degree of the road sections,and the importance degree of the road sections is obtained,so that the key sections in the urban road network can be identified.(3)in the research of traffic state identification,this paper selects the traffic flow,saturation and average travel speed as the evaluation indexes of traffic state identification,by using Kmeans Clustering Method and FCM Clustering Method,the traffic status of different grade urban roads is analyzed,five kinds of traffic states are obtained: Serious Traffic Jam,moderate traffic jam,mild traffic jam,unblocked traffic and very unblocked traffic.In addition,the clustering results of K-means Clustering Method and FCM clustering method are compared,and the clustering center of traffic state is determined,and the clustering results of the two methods are similar.(4)in the research of traffic state prediction,in order to improve the prediction precision of traffic parameters and traffic state,N-BEATS neural network model is introduced,combined with FCM traffic state discriminant model,n-BEATS-FCM traffic state prediction model is proposed.In this study,the prediction effect of N-BEATS model is compared with the traditional RNN,LSTM and RNN-mlp neural network models.The results show that N-BEATS model has less prediction error.The speed prediction error of N-BEATS model is 6.10%,and the traffic error is 7.78%.The accuracy of traffic prediction is 83.8%.The training and testing speed of the model is fast,which verifies the good performance of N-BEATS-FCM.In this paper,a set of methods for identifying critical sections of urban road network,distinguishing and forecasting road traffic states are constructed.The research results can be used to monitor urban traffic based on license plate recognition data,and provide effective information for traffic management departments to support the implementation of decisionmaking and Operation Management Strategy.