Research On Urban Road Traffic State Prediction Based On Floating Vehicle Data

Recently the urban road traffic congestion is increasingly serious,affecting people's daily life.Especially in the morning peak and late peak and other period hours,road congestion will be extremely serious.However,at this time not all roads are in a state of congestion,through the prediction of urban road traffic conditions,you can master the traffic information,to maximize the efficiency of the use of urban road resources,to ease the congestion on the road traffic pressure.Therefore,the intelligent transportation system which includes a variety of advanced technologies has gradually become the effective method to solve the problem of urban traffic.Urban road traffic state forecasting in intelligent transportation system is the key to traffic control and traffic guidance.After summarizing a large number of literature,this paper presents a urban road traffic state forecasting model,which includes urban road speed prediction as well as urban road traffic state discrimination.By comparing the relevant model for traffic flow forecasting,this paper establishes a BP neural network based on improved particle swarm optimization for urban road speed forecasting model.The concept of "crossing"and "variation" of the genetic algorithm are introduced into the classic particle swarm algorithm,this can overcome problem such as it is easier to fall into local optimal value.With improved particle swarm optimization algorithm,the classic BP neural network's weight matrix and the threshold value are improved,makes the model own the higher accuracy.In view of urban road traffic state discrimination,it is established based on fuzzy c-means clustering algorithm of traffic state discrimination model in this paper.It can reflect sample and relationship of every clustering center,distinguish classification.In this paper,the actual floating vehicle data of some taxis in Dalian are pretreated and applied to the urban road traffic forecast model.The experimental results show that the forecasting model of urban road traffic can better predict the traffic status of urban roads and have some reference significance for traffic guidance and control,so that the limited urban road network resources can play a bigger effect.