Study On Short-term Traffic Flow Prediction Of Urban Expressway Based On Spatio-temporal Characteristics

As the highest-grade road in the urban transportation system,the driving quality of urban expressways directly affects the operational efficiency of the entire urban road network.However,the rapid increase in the number of motor vehicles and the demand for people’s travel has caused traffic congestion on the expressways to become more and more serious.With the development of intelligent transportation systems,there are new ways and methods to alleviate traffic congestion on expressways and improve urban operating efficiency.In the intelligent transportation sysytems,predicting the future traffic flow parameters based on real-time traffic information is the basical and key work.Therefore,in order to alleviate traffic congestion with the help of intelligent transportation systems,it is significative to predict real-time and accurate traffic flow parameters in real time.By combing,inducing and summarizing the domestic and foreign research contents on short-term traffic flow prediction of urban roads,this paper concludes that although many valuable research results have been achieved in the traffic flow prediction on a single section,the data acquisition and prediction methods are still need to be improved in some parts.On this basis,this paper focuses on the prediction of the average travel speed on the expressway by considering the time and space characteristics of traffic flow parameters.First of all,this article introduces the method and main steps of obtaining data in detail.Aiming at the related problems when using Python to crawl the average travel speed data,a solution for data cleaning,repair and noise reduction is proposed,which lays the data foundation for improving the prediction accuracy of the model below.Then,from the perspective of traffic flow theory,a brief introduction to the basic parameters of the traffic flow and the relationship model between the parameters is introduced.Based on the collected traffic flow data,the macroscopic characteristics of the traffic flow parameters on the Xi’an Expressway are analyzed and the relationship between different parameters are fiited as function.In order to gain a deeper understanding of the time and space characteristics of traffic flow parameters,this paper combines the collected average travel speed data to analyze the traffic flow in three dimensions: time dimension,space dimension,and time series,which lays the foundation for constructing prediction models later.Based on the wavelet neural network model,considering the temporal and spatial characteristics of the traffic flow on the urban expressway,four different state vectors are constructed,to establish the short-term traffic flow prediction model of the urban expressway.In order to test the applicability of wavelet neural network in the prediction of short-term traffic flow on a single cross-section and the accuracy of the prediction model constructed,this paper used the average travel speed data collected on Xi’an Second Ring Expressway to verify the prediction model.The results show that the wavelet neural network is suitable for short-term traffic flow prediction on a single cross-section.The accuracy of traffic flow predicting by wavelet neural network is higher than which predicting by back propagation neural network,and its prediction accuracy is more than 90%;in conclusion comprehensive consideration the prediction model of traffic flow time and space characteristics can effectively improve the prediction accuracy of the model.