ETC Data Based Identification Of Vehicles Entering The Service Area And Traffic Flow Prediction

Obtaining the traffic flow in the expressway service area can enable administrators of expressway service areas to reasonably allocate the resources,and travelers can enjoy better services.There are not vehicle detection equipments in most service areas now so that the traffic parameters of the service area can’t be obtained conveniently and accurately.Therefore,how to use the existing expressway equipment to obtain the traffic flow situation of the expressway service area is of great significance both to the facility layout of the service area and to the improvement of the service capacity of the service area.Based on the analysis of vehicle speed distribution in the service area,the paper puts forward the discriminant method of vehicle entering the service area based on the speed division.Based on the analysis of the consistency of vehicle speed on expressway,a discriminant method of vehicle entering the service area based on variational inference gaussian mixture clustering is proposed.Finally,based on ETC data,the factors affecting the traffic flow in the service area are analyzed and the vehicle flow prediction method at the entrance of the service area is established.The main contents include:(1)A method of discriminating vehicle entering service area based on the speed division.Vehicle detectors are not installed in most service areas,so the vehicle information at the entrance of the service area can’t be obtained.This paper uses the ETC data analysis of vehicles in service area road speed distribution law.This article uses the ETC data analysis of vehicles in service area road speed distribution law.And then it puts forward the GMM-EM algorithm to find the probability of passenger cars and trucks in the service area road speed curve.Considering that the EM algorithm is sensitive to the initial value,kmeans is used to initialize the parameters of EM algorithm.And the speed threshold of the bus and truck driving into the service area is solved by the simulated annealing algorithm.The validity of this method is verified by using the historical data of Chongqing highway.(2)A discriminant method of vehicle entering service area based on variational inference gaussian mixture clustering.It is one-sided to judge whether a vehicle enters the service area only with the speed threshold.Based on the analysis,this paper concluded that the vehicle speed was consistent.So it calculated the ratio of the driving speed of vehicles on the road sections in the service area to the driving speed on the upper/lower sections respectively.And it classified the vehicles according to the driving path,and established the variational inference Gaussian mixture clustering algorithm for each category,respectively to discriminate whether the vehicle drove into the service area.This paper uses the historical data of Chongqing highway to carry out the experimental verification and gets the accurate value and the applicable range of each input combination.(3)Forecasting method of vehicle flow at the entrance of service area based on ETC toll data.At present,there is no special method to forecast the traffic flow at the entrance of service area.In this paper,the previous two methods are used to estimate the historical incoming traffic flow in the service area.After analyzing the influencing factors of the vehicle flow in the service area,the CNN-BILSTM-ATTENTION model is built to predict the vehicle flow at the entrance of the service area.Based on the data of a expressway in Chongqing,the efficiency and the superiority of the proposed prediction model is proved by comparing with other prediction models.