BIT* -based Path Planning for Remotely Operated Maritime Autonomous Surface Ships Under Dynamic Obstacle Uncertainty

This study proposes a BIT* (Batch Informed Trees)-based path-planning approach for MASS (Maritime Autonomous Surface Ships) that accounts for dynamic obstacle uncertainties. In congested waters, safe navigation during remotely operated MASS requires reliable avoidance of dynamic obstacles in complex port environments. However, sensor measurement errors and communication delays introduce positional uncertainty that degrades the safety and reliability of planned paths. To address this challenge, collision avoidance and return-to-global-path planning are formulated as a unified optimization problem within the BIT* framework, enabling consistent decision-making under uncertainty conditions. By incorporating sampling-based probabilistic collision risk into the BIT* cost function, the search is biased toward regions with lower collision probability. COLREGs (Conventions on the International Regulations for Preventing Collisions at Sea) are also integrated to induce rule-compliant maneuvers in head-on and crossing situations. Numerical simulations demonstrate that the proposed approach effectively generates safe and reliable paths under varying levels of positional uncertainty.