Quantifying Good Seamanship: A Framework for Risk- Aware Trajectory Planning in Autonomous Vessel Navigation

Autonomous vessel navigation plays a vital role in enhancing maritime safety and operational efficiency, particularly in congested and complex waterways. Navigators should always display “good seamanship”. This requires the ability to predict the future evolution of scenarios, assess risks, and plan several steps ahead. However, quantifying good seamanship remains challenging due to the limited ship maneuverability, multi-vessel encounters and near ground scenario. We have developed a framework for quantifying the good seamanship score based on the collision and grounding risk that takes into account the vessel kinodynamics constraint and the uncertainty of target vessel speed. This framework helps evaluate seafarers' actions in avoiding collisions and groundings. Utilizing a quantified measure, a riskaware trajectory planning algorithm has been developed, which uses sampling-based methods to produce a safe and efficient trajectory, which is tested in various scenarios. Future efforts will address hydrodynamic effects such as wind and waves, consider target vessel future action, and explore advanced machine learning techniques to further improve the computation speed in scoring the good seamanship and generating the trajectory in complex scenario.

Research Questions:

- How can we define and quantify a reliable risk or good seamanship metric that accounts for all situations, including uncertainties, maneuverability limitations, different mission and vessel types, and hydrodynamic effects?

- Given these factors, how can we develop a real-time algorithm that generates the optimal trajectory for an autonomous vessel?