Optimizing the anti-collision decisions between multiple ships based on predictive risk evaluations

Ship collisions can cause great losses of lives and property and brings enormous negative impacts to the maritime environment. To improve the intelligence and autonomy in assisting ships to make safe and quick decisions to avoid potential collisions, this paper proposes an optimization method for multiple ships to avoid collisions, based on predictive risk evaluation. Firstly, each ship makes trajectory predictions based on ship maneuverability model, and the inter-ship collision risk parameters calculation within a prediction horizon are calculated. To ensure that the ships choose suitable rudder angles to finish the collision avoidance process in an efficient way, an optimization problem is formulated using the knowledge with respect to the inter-ship collision risks and solved with a commercial solver. Preliminary results show that the proposed method can provide ships with the rudder angles they should choose and the corresponding rudder steering time to keep enough safe distance with the other ships.