Research On Vessel Traffic State Identification Considering Speed Dispersion Characteristics

Vessel traffic flow is one of the important research directions of maritime traffic engineering.Through the study of vessel traffic flow,the behavior characteristics and rules of ships can be analyzed quantitatively or qualitatively,which can provide effective means and methods for identifying sea navigation risks and improving the efficiency of vessel traffic.Existing studies have used traffic flow parameters to describe ship traffic flow characteristics from a macro and overall perspective,without considering the interaction between individual behavior differences and macro overall performance.Therefore,the actual operating conditions of vessel traffic flow cannot be comprehensively and accurately described.It is of great significance for the study of vessel traffic state and mechanism to introduce the study of speed dispersion characteristics and understand the individual differences in vessel traffic flow.Based on the traditional land,from the macroscopic angle,on the basis of description of vessel traffic flow was introduced into the speed dispersion index,analysis of speed dispersion characteristics with the relationship between the macroscopic traffic flow parameters,to individual differences in behaviour with macro creates a link between overall performance,from the perspective of qualitative research for the running mechanism of vessel traffic flow;On the basis of the above,through data mining,from a quantitative point of view,the evaluation and identification model of vessel traffic state is established.The main contents are as follows:(1)Study on speed disperision of vessel traffic flow.Firstly,based on the definition and summary of speed dispersion,the definition of vessel traffic flow speed dispersion is proposed,the distribution of speed dispersion on the basic diagram of traffic flow is studied,and the relationship between speed dispersion and macro traffic parameters is analyzed.(2)Research on speed traffic state evaluation.In view of the limitations of macro parameters used in traditional research on traffic flow condition,speed dispersion is introduced as the evaluation index,and Fuzzy-C-Mean(FCM)clustering algorithm is used to evaluate traffic flow condition.To solve the problem that the initial clustering center of Fuzzy-C-Mean is random and easy to fall into local optimization,an improved K-mean algorithm is introduced for optimization,a vessel traffic clustering algorithm with improved FCM is proposed.(3)Research on the status identification of vessel traffic flow.According to the result of vessel traffic flow condition classification,the application of generalized regression neural network(GRNN)and probabilistic neural network(PNN)to learn all states of traffic condition data set,is established based on the GRNN and PNN vessel traffic state identification model,based on the results of the identification.Through repeated training and result verification,the accuracy rate is more than 95%,which proves the feasibility and effectiveness of the model.To sum up,this dissertation describes the characteristics of vessel traffic flow more comprehensively and accurately through the understanding of speed dispersion characteristics in the operation of vessel traffic flow.On this basis,the vessel traffic status clustering and identification has done a relatively sufficient study,which can provide some reference value for the related research.