Dynamic Route Selection Problem Based On Short-term Traffic Flow Forecasting Research

With the continuous development of various artificial intelligence science and technology,the ITS(Intelligent Transportation System)has been integrated into the daily lives of ordinary people.As a significant development direction of the future transport system,the ITS combines the advanced science,technologies,artificial intelligence and other ideas effectively,and employs them in the transport system,then establish a real-time,accurate and efficient comprehensive transportation and management system,which could play a more comprehensive role on a higher level.The dynamic routing selection of urban transport is one of the important links in the ITS,it has also become a hot issue in academic research in recent years.Heavy traffic congestion is a serious problem which always cannot be improved in China,and the “Navigation Congestion” is a new situation appeared recently.In the light of these problems,this paper starts with the dynamic routing selection of urban transport network,designs a dynamic path selection scheme based on short-term traffic flow forecast.The holistic thought of this paper is introduced as follow.Firstly,produce short-term(5 minutes in the future)traffic flow forecast of traffic condition based on real-time information on each road in urban transport network,then reflects the variation of traffic conditions.Secondly,convert the forecast information into the estimated average journey time of each road by the speed-time conversion formula,then identify the value of road right of each road.Thirdly,determine the optimal path at present based on the improved algorithm of traditional Dijkstra algorithm.This scheme will continuously circulate the short-term forecast and routing selection of the traffic flow until the traffic participants arrive at their destination.Then obtain the balance of traffic conditions in urban transport network,improve the situation of traffic congestion and avoid the appearance of "Navigation Congestion" and other bad road conditions finally.On the basis of summarizing the research status of short-term traffic flow forecast and routing selection in urban traffic network at home and abroad,by analyzing and filtrating existing model and algorithm,based on the influence of traffic flow forecast information on the efficiency of the whole urban transport dynamic routing selection,because of the high accuracy requirement of the result of the forecast,a speed forecast model of wavelet neural network is proposed in this paper.However,according to some basic characteristics ofdynamic routing algorithm and the determination of some key factors,the algorithm which is chosen for the dynamic optimal path is required to have higher search efficiency and less time complexity.Therefore,in this paper,the algorithm based on the traditional Dijkstra algorithm with improvement of time complexity is chosen to solve the dynamic optimal path.In order to verify the effectiveness of the dynamic routing selection scheme designed in this paper based on short-term traffic flow forecast,a part of road network near Hefei Pearl Square was selected as an example and introduction of the basic road traffic conditions and simulation platform is proposed,and finally verifies the feasibility of the scheme designed in this paper.