Study On Regional Ship’s Arrival Model Based On AIS Data

The research of marine traffic simulation technology involves a large scale complex system,in which the regional ship’s arrival model plays an important part. According to the theory oftraffic flow, the regional ship’s arrival model can be described by the ship arrival pattern andtime-interval between the ships pattern. In order to fix the model’s characteristic parameter, thedata is obtained by the marine traffic survey. However, traditional marine traffic survey’smethods are time-consuming and trouble, which is not convenient to research. Now an importantnavigational aid–Automatic Identification System (AIS) has been made compulsory for mostships around the world by the International Maritime Organization (IMO). We can get abundantship’s dynamic and static information from the AIS message. Based on AIS data, using the datamining technology and mathematical statistics, the paper builds the regional ship’s arrival model.And then the validity of the model is verified by the AIS data. The main works are as follows:1) The numerical characteristics of the marine traffic are obtained by Online AnalyticalProcessing (OLAP). The OLAP of ship traffic is designed by the historical AIS database anddata warehouse, which brings the numerical characteristics of arrival model.2) The time-interval between the ships is established, and then to realize for the efficientmining algorithm. According to the characteristics of AIS, the moment of the ship passingthrough a certain survey line is obtained by the Time Correction Mode of vessels through thesurvey line (TCM) algorithm and Straddle Test (ST) algorithm. The efficiency and effectivenessalgorithm is finally confirmed by comparing the TCM algorithm and ST algorithm.3) Using marine traffic engineering, the ship’s arrival model for port area is set up. Andthen the validity of the model is verified by the AIS data. Taking the Xiamen port, Qingdao port,Lianyungang port and Yangtze River channel as examples, their characteristic parameters ofship’s arrival model are obtained by AIS data. And the generality and individuality of ship’sarrival pattern within the regional area is analyzed for detailed knowledge.The result shows those ship’s patterns are mainly subject to a negative binomial distribution,Poisson distribution and binomial distribution in different parameters, and then the densitydistribution of time-interval between the ships is subject to a GAMMA distribution in differentparameters, which fits better than those frequently-used distribution models. The study results ofregional ship’s arrival model can provide theoretical basis for maritime authorities in protectingthe marine navigation environment, reducing ship accidents and improving the efficiency oftraffic organization. This model can also be used in forecasting and some other ways. In additon, the model plays an important role in further research on marine traffic in the event of more rapiddevelopment of port economy and more complicated traffic flow.