Collision Risk Assessment Techniques for INF Ship Using AIS Data

We propose techniques of arranging traffic information of a sea area using AIS data and then estimating conditional collision frequency for a ship passing the area. Then we perform a case study to show applicability to collision risk assessment. For two assumed routes of the most crowded coastal waters in Japan, route-C (close to the land) and route-F (far out at sea), we estimate collision frequencies against all ships, tankers and very large ships, and then discuss route selection in terms of INF ship collision risk. One of the findings is that average collision frequency for route-C, which is considered as a maximum measure of Japan coast, is at least a factor of 3 lower than the estimate of the SeaRAM study. The difference is caused by uncertain target population assumed in SeaRAM. This study proves that AIS data allows us to know accurate ship population and then estimate realistic collision frequency under a certain condition such as route, ship group to be collided, etc. This will open a new vista for a risk-based routing strategy of INF ship considering various relevant factors.