Font Size: a A A

Study On Vessel Traffic Flow Prediction And Simulation Based On AIS Data

Posted on:2017-06-07Degree:MasterType:Thesis
Country:ChinaCandidate:X Y ZhuFull Text:PDF
GTID:2382330566953003Subject:Traffic Information Engineering & Control
Abstract/Summary:PDF Full Text Request
Maritime transportation has always been the dominant support of global trade,and with the growth in world trade,there is a massive increase in shipping.And the ports are facing growing safety and security issues with higher traffic flow,the research of vessel traffic flow prediction and simulation can provide reference for the traffic management of the port water area.Basedon thestudy of modeling and predictive methods of neural network,a prediction model of vessel traffic flow was proposed in thisthesis,and the port service model wasestablished by system simulation based on the analyze of the AIS(AutomaticIdentificationSystem)data.The research contents are as follows.Firstly thethesisanalyzed the basic knowledge of vessel traffic flow theory and vessel traffic flow prediction theory,and the indexes for evaluation of the vessel flow prediction weregiven.Then the thesis established a vessel flow prediction model based on the study of the neural network modeling and wavelet analysis,which takes advantage of the time-frequency localization of wave transform and the self-learning ability of neural networks.In order to improve the accuracy and the convergence speed of prediction model,the Particle Swarm Optimization(PSO)algorithm was introduced.While training the Wavelet Neural Network,using the PSO algorithm to optimize the connection weights,scale factors and shift factors of the WNN model,established the PSO-WNN prediction model.And then forecasted the vessel traffic flow of Xiazhimen channel with WNN prediction model and PSO-WNN prediction model,and the historical traffic volume data of Xiazhimen channel can be obtained through AIS data.And the results showed that both WNN prediction model and PSO-WNN prediction model,at a certain level,could predict the trend of the traffic flow.Compared the outcomes of the two prediction models,the results indicated that thePSO-WNN prediction model had higher accuracy and faster convergence speed.The queuing theory was used to set up a mathematical model to describe the port service system.For the arrivals were Poisson distributed and the service time wasexponentially distributed,the port service system could be described by M/ M/ c model,and analyzed the port service system using system simulation.
Keywords/Search Tags:AIS Data, Traffic Flow Prediction, Wavelet Neural Network, Queuing Theory, Simulation
PDF Full Text Request
Related items