Study On Urban Road Traffic Flow Short-term Prediction And Its Application

In recent years,with the increase of car ownership in China,traffic congestion has become a prevailing problem in urban traffic.Traffic signal control strategy optimization is an effective solution to alleviate traffic congestion with low cost and high efficiency.Traffic flow short-term prediction and traffic congestion identification are the key and basis of traffic signal control strategy.Traffic control zoning is the guarantee of efficient,safe and reliable operation of traffic prediction and traffic control system.Based on the theories of complex network,Bayesian network,quantile regression,probability theory and iterative search,feedback control,with the purpose of low-cost and high-efficacy to ease traffic congestion,this paper is trying to make an inductive summary of traffic control partition method,short-term traffic flow prediction,traffic prediction and recognition in time of congestion,traffic signal control strategy optimization,and in view of the existing problems and difficulties in these studies,some solutions are proposed.The main research work and innovations are as follows:1)Taking as indicators static data such as topology structure of urban road network and dynamic data such as real-time traffic flow,which are easy to collect,a subdivision method combining static and dynamic data is proposed.When the traffic in the road network is in free and fluent flow,the static traffic partition of the road network is only carried out according to the static traffic data such as the network topology structure,so as to improve the operation efficiency of the traffic control system in free flow.At the same time,static data is used to solve the problems when some prior parameters are difficult to determine in initial partition.When the traffic flow of any section of the road network is large and the average density of traffic flow reaches the threshold,dynamic division is carried out on the basis of static division to alleviate traffic congestion in the control sub-area.It has been testified that this method can be reasonably partitioned in both congestion and non-congestion periods,and the partitioning effect is better than other methods.2)In the process of traffic area division,this paper proposes to use the intersections as the clustering center of the initial division of traffic region.Therefore,in the previous process of traffic area division,the number of traffic zones was difficult to determine,the homogeneity of road network structure difficult to measure,and the size of traffic sub-areas difficult to control.Now they are solved.3)In order to solve the two difficulties in short-term traffic flow prediction,that is,the temporal and spatial correlation of traffic flow and the change of traffic flow caused by abnormal conditions,this paper proposes a hybrid prediction method based on quantile regression and Bayesian inference network.This prediction method can solve the problems caused by the temporal and spatial correlation of traffic flow and the change of traffic flow under abnormal conditions.It is proved that the prediction accuracy of this method is better than other methods when real-time traffic flow is accurately known and collected,and the prediction accuracy of abnormal traffic flow changes is much higher than other methods.4)In the paper expounds the application of traffic flow short-term prediction in traffic congestion identification and traffic accident identification.At the same time,the traffic signal adaptive predictive feedback control and traffic signal fixed feedforward feedback control are compared and analyzed,and the optimization effect of traffic flow short-term prediction on traffic signal control system is proved.In the paper mainly focuses on the short-term prediction of traffic flow.The traffic zoning method is proposed to ensure the efficient operation of short-term traffic flow prediction and traffic signal control.At the same time,the application of traffic flow short-term prediction in traffic congestion identification,traffic accident identification and traffic signal control strategy is studied.