A cellular automation model for convoy traffic in Arctic waters

Marine transportation in arctic waters involves risk due to the presence of ice, rapid weather changes, reduced visibility, and remoteness. Icebreaker assistance and ice-convoys are used to facilitate the safe navigation of commercial vessels transiting Arctic waters. Collision between individual ships, or a ship and an icebreaker, during a convoy is a potential danger. To study such risk, this paper presents an updated Nagel-Schrekenberg (NaSch) model of the marine convoy traffic integrated with a Bayesian Network (BN) based probabilistic approach. This approach has been used to predict the maximum waterway density for the safe flow of traffic, and the collision probability during a convoy. The approach is tested here on a convoy in the Vilkitskii strait. The BN model proposed here could assist in assessing the failure probabilities of the causal factors and their contribution to collision likelihood. It may also be useful in developing a collision monitoring system that provides a real-time estimate of collision probability. The estimated probability could be used as an early warning for potential collisions, thus enhancing safe navigation and safety at sea.