In order to develop safe navigation environment, it is important to know the characteristics of ship traffic flow. Although there are some correlations between parameters of ship traffic data such as position, velocity, and ship size, they have been dealt individually in sea traffic simulation in past researches. To carry out more realistic ship traffic simulation, it might be necessary to generate ship traffic data taking into account the correlations.

In this study, authors developed a method of characterization and generation of ship traffic flow by using Principal Component Analysis. By applying the method to AIS data in the Kanmon Passage, the trend of the generated ship traffic data which is same as characteristics of AIS data was shown. Authors also developed a method of simulation of evaluating encounter probability between ships. Encounter conditions of ships with collision risks could be extracted, so that the encounter probability could be calculated. To show the validity of the method, the computed probability was compared to an existing estimated collision probability.