Conflict detection method based on dynamic ship domain model for visualization of collision risk Hot-Spots

Ship collisions have occurred frequently in the past few decades due to the significant increase in traffic volume. In view of their severe consequences, it is of practical significance to identify the ship encounter scenarios with collision potentials for the waters with high traffic density. In this paper, a systematic methodology is proposed to detect possible collision scenarios and identify the distributions of collision risk Hot-Spots in a given area. A dynamic ship domain model that considers the preferences of navigators with respect to the distance passing with other ships before and after the collision avoidance action is first built as the basis of conflict detection modelling. Then a ship domain-based conflict severity model is established to evaluate the risk levels of ships being involved in dangerous encounters. The spatial-temporal pattern mining methods are subsequently extended to identify collision risk Hot-Spots and analyze the collision risk cluster features with time of day. A case study is finally performed using the real AIS data in Ningbo-Zhoushan Port area. Results indicate that combining spatial-temporal pattern mining with collision candidate detection is a useful means to identify where the Hot-Spots are and how the risk is distributed in terms of space and time.