The Design And Realization Of The Ship's Trajectory Clustering Prototype System Based On AIS

Nowadays,ships loaded with equipment of Automatic Identification System(AIS)are increasingly growing in number.Trajectory data located by AIS equipment is only able to reflect the information such as current longitude and latitude,sampling instant,instantaneous velocity and ship course;however,it cannot reflect the sailing characteristic information which is extremely vital in marine traffic planning,management and control.Therefore,the ship trajectory clustering prototype system is designed and realized with the Java language through Eclipse software.The main purpose of this thesis is to mine and analyze the AIS data information,and then to obtain the ship's behavior patterns.Firstly,the trajectory clustering methods and the AIS data mining applications have been surveyed in the thesis.Secondly,based on the theory of data mining and the relevant navigation technology,the overall design of the ship's trajectory clustering prototype system is given.The prototype system is designed by hierarchical and modular design principle,and the prototype system is divided into three layers,i.e.,data source layer,data processing and mining layer and user interface layer from the bottom to the top.The AIS data source layer includes AIS database module,which is to store the raw AIS data and the AIS data after preprocessing.AIS database module is developed to provide data query,data deletion,data insertion and other functions for the upper layers.Data processing and mining layer includes AIS data preprocessing module and AIS data mining module.The function of AIS data preprocessing module is to remove the AIS redundant data and abnormal AIS data,to filter drifting data,to fill the missing AIS data,and to separate the AIS data.AIS data mining module includes two sub-modules.One sub-module is to find the ship's regions of Interest and to calculate the interest degree of regions of interest based on the clustering algorithm to obtain the ship's mooring point.The other sub-module is to cluster the segment data of the sub-trajectory,and to extract representative trajectory in cluster and further to discover the hot ship route.User interface layer implements the user interface of both the data or parameter input and output of data mining results.Finally,we conduct the experiment to test the functions and performance of the prototype system.Real AIS data provided by the maritime bureau are used in the experiment.The test results show that all the modules can satisfy the design requirement.In the experiments,we analyze the effect of parameter values on the clustering result in the data mining stage and give the appropriate experience parameters,which can provide the reference for the real operation of the prototype system.Moreover,the case studies results have demonstrated that our designed AIS-based ship trajectory clustering prototype system can find the ship interest area and the hot ship route effectively and efficiently.