Research On Traffic State Discrimination And Risk Analysis Based On GIS

The increase in car ownership has led to frequent traffic accidents,which has caused widespread concern.aroused people’s extensive attention.The congestion status of urban roads is closely related to many indicators such as accident rate and accident severity,caused by traffic jams and at the same time,a vicious cycle of accidents have increased congestion.accident rate,accident severity and other indicators.Traffic jams lead to accidents,and accidents exacerbate congestion.With the rapid development of intelligent transportation system,real-time visual display,which of the accident location,road traffic conditions and vehicle risk status,has theoretical and practical significance for subsequent traffic control,traffic guidance,risk control.traffic congestion at planning level and the proposed of the risk prevention and control methods.Therefore,from the perspective of intelligent transportation,this paper made an in-depth study on the theory of traffic state discrimination and accident cause analysis,and a traffic state discrimination method based on the fuzzy C-means(FCM)clustering and support vector machine(SVM)and an accident causal analysis method based on the principal component analysis(PCA)and equivalent fuzzy clustering(FCA)were proposed,and a road and vehicle monitoring system was developed,which realized the visual display of traffic and risk state discrimination.The main work includes the following contents:Firstly,the clustering method related to traffic state discrimination and dangerous state identification is analyzed.The principle,advantages and disadvantages of fuzzy clustering and support vector machine are studied.Then,the traffic state discrimination method based on FCM-SVM is proposed.The accuracy of the algorithm was verified by the real data.Secondly,an accident causal analysis method based on the the PCA-FCA was proposed.The applicability of the algorithm for vehicle dangerous state identification is verified by the actual traffic accident statistics.Thirdly,on the basis of analyzing the data structure of the database and under the principle of following the basic idea of GeoDatabase data organization,the related steps of data view modeling were simplified,and the database storage scheme was developed.By designing the data model of database,the design of the database of the traffic state and vehicle risk discrimination was realized.Fourth,the traffic state discrimination and danger identification system was developed.The data acquisition and storage of the GPS data,monitoring center data and intelligent vehicle monitoring equipment were described in detail,and the development process and interface design of each functional module were explained,including the monitoring and display of accident location,the discrimination and display of traffic status and the dangerous vehicles.