Objective To address the complex cooperative collision avoidance problem among multiple intelligent ships in intersection waters, this paper proposes an enhanced large language models(LLMs)-driven decision-making method for multi-ship cooperative collision avoidance in intersection waters.
Method By analyzing the navigational characteristics of intersection waters, this study formulates the multi-ship cooperative collision-avoidance problem as a Partially Observable Markov Decision Process (POMDP), thereby providing a formal mathematical foundation for subsequent decision making. The cooperative process is decomposed into four causally linked modules—state perception, intent sharing, conflict coordination, and avoidance decision—to structure both information flow and reasoning. Guided by human seafarers’ embodied practices, a novel central-distributed dual-layer architecture is proposed. At the central layer, an LLM-based coordinator aggregates multi-ship situation, applicable navigation rules, and quantifies conflict severity to infer passage-priority sequences. At the distributed layer, individual ship agents are empowered by LLMs together with chain-of-thought prompt engineering to perform progressive, stepwise reasoning. These agents synthesize structured scene descriptions, coordination directives and retrieved navigation experience to generate executable avoidance maneuvers and accompanying semantic explanations. To mitigate known weaknesses of LLMs in precise numerical computation and continual learning, and to suppress potential hallucinations, the architecture integrates two complementary augmentation mechanisms. A lightweight mathematical engine is invoked to update kinematic states and compute deterministic conflict metrics, supplying rigorous quantitative inputs to the reasoning pipeline. A retrieval-augmented generation (RAG) navigation knowledge base combines a static corpus of rules with a dynamic repository of historical scene–decision–evaluation tuples, enabling case-based grounding and continuous learning from past interactions. By embedding formal computation and evidence-backing into the LLM reasoning loop, the proposed framework preserves the interpretive and interactive strengths of large models while ensuring verifiable, rule-compliant, and practically executable collision-avoidance decisions in complex intersection waters.
Results Simulation experiments demonstrate that the proposed enhanced LLM-driven method implemented in DeepSeek-v3 achieves safe and efficient cooperative collision avoidance in typical intersection scenarios involving two, three, and four ships. The system maintains a minimum maneuvering speed of over 3 knots throughout and ensures a safety margin exceeding twice the ship length.
Conclusion This method advances the engineering application of LLMs in maritime decision-making and provides a new pathway for realizing highly autonomous shipboard artificial intelligence in complex operational environments.
DownLoad: