This research introduces a novel LLM-based decision-making system for autonomous surface vehicles that complies with maritime collision regulations (COLREGs).

• Transforms complex maritime rules into human-like decision processes for autonomous vessels
• Enables ASVs to interpret and apply collision regulations designed for human operators
• Demonstrates effective obstacle avoidance while maintaining regulatory compliance
• Represents a significant advancement in explainable AI for maritime applications

Engineering Impact: This approach bridges the gap between rigid algorithmic systems and human maritime expertise, potentially accelerating the practical deployment of autonomous vessels in complex real-world environments with regulatory requirements.

Large Language Model-based Decision-making for COLREGs and the Control of Autonomous Surface Vehicles