Research On The Traffic Flow State Evolution And Identification Method Of Regional Road Network Based On Taxi GPS Data

In order to solve the problem of the growth rate of vehicles owned is much higher than the traffic infrastructure construction,use ITS technology to solve the problem of traffic is an inevitable trend.And it is the basis of ITS technology to accurate identifies the real-time and future traffic flow.Urban traffic system is a complex system,which has strong randomness and dynamics,and also has strong regularity and connection.According to this,this paper takes some regional of Xi'an as an example,studies the traffic flow state identify method based on the GPS large data of the taxi,and excavates the evolution characteristics of the traffic flow state,so as to grasp the real-time and accurate traffic flow operation state and identify recurrent and incident traffic congestion.And it is the basis for regional traffic management and control,and it can ensure the urban regional traffic safety and smooth operation.Therefore,the following aspects are discussed in my paper:(1)Research on traffic parameter estimation methodIn this paper,the GPS data of the taxi is used as the research data,and the traffic flow state can't be directly and accurately determined by the information such as the speed,GPS time,latitude and longitude coordinates,direction,vehicle state and data validity provided by the original data.So use this information to estimate the traffic parameters,and thus indirectly obtain more effective information.Under the precondition of pre-processing and map matching,the traffic parameter estimation model such as the average speed of the road,the average delay of the road and intersection is constructed,and it is the foundation of the following research works.(2)Research on traffic flow state identification based on SVMTraffic flow state recognition model based on SVM binary tree classification algorithm is constructed.And three traffic Parameter as input data,through several training calibration model parameters.The experimental results show that the model can solves the problem of nonlinear and high dimension,and realizes the accurate identify the traffic flow state in urban regional network.(3)Research on traffic flow state evolution characteristics analysis methodThe traffic flow state identification model is used to identify the traffic flow state of the urban regional road network.Through constructing the traffic flow state migration network model based on the Markov model and the traffic flow state evolution characteristic analysis model,the migration characteristics,stability,preference,activity,active time,skip migration state,congestion rate,space-time distribution and other characteristics of urban regional road network traffic flow state are excavated,so as to more in-depth understanding of the overall operation of urban regional.It is the basis of urban regional for the development of traffic management and control programs.(4)Application of traffic flow state identification and evolution characteristicsTraffic parameter estimation method,traffic flow state identification method and traffic flow state evolution characteristics analysis method were combined to verify the accuracy,reliability and practicability of the model based on the some regional road network,taxi GPS data and video data in Xi'an hi-tech industries development zone,and analyzes the evolution process and characteristics of traffic flow in regional road network,identifies recurrent and incident traffic congestion.