Knowledge-Enhanced Large Language Model-Driven Motion Scenario Generation Method for Human-Robot Collaborative Robotic Arms

In human-robot collaborative manufacturing under intelligent manufacturing paradigms, this paper proposes an integrated framework that synergizes highly interpretable domain-specific knowledge graphs with general-purpose large language models (LLMs). The primary objective of the framework is to mitigate hallucination risks and domain knowledge gaps inherent in LLMs for industrial applications. It does this by constructing a structured knowledge repository for computer case assembly domain knowledge graphs through a pattern layer. The Lang Chain framework enables high-precision knowledge retrieval, incorporating self-evaluation mechanisms. It facilitates natural language-based user requirement analysis, terminology correction, and rule inference. The study designs a domain-specific knowledge graph schema layer to establish an accurate knowledge repository. It leverages general large language models' semantic comprehension and reasoning capabilities to enable user demand analysis, terminology correction, initial solution generation, domain rule inference, and human-machine interpretable plan formulation. The experimental results confirm that this hybrid approach preserves LLMs' semantic comprehension strengths while enhancing decision-making accuracy and interpretability in industrial scenarios, offering technical foundations for knowledge services in human-robot collaborative manufacturing systems.