Uncertainty-Aware Decision Making for Safe Navigation of Autonomous Ships

Abstract

During ship encounters, uncertainty surrounding other vessels’ intentions and future trajectories can complicate situation assessment. When encountering an erratic vessel, navigators will take precautionary actions to avoid any risk of collision. This paper introduces a new decision-making algorithm tailored to address uncertainty in vessel encounters at sea. The proposed method achieves this by defining the risk of a future scenario as the combination of two factors: the probability of collision and the magnitude of the uncertainty. As this method does not assume deterministic future trajectories for either of the vessels, it can therefore be applied to any pair of vessels, and it allows the use for anticipation of future behaviours of other vessels in a scene. The method is verified through a simulation study.