Focusing on the integration of Large Language Models (LLMs) in robotics, this research examines their role in path planning and providing linguistic explanations for autonomous intelligent robots. LLMs have demonstrated superior performance in natural...
Focusing on the integration of Large Language Models (LLMs) in robotics, this research examines their role in path planning and providing linguistic explanations for autonomous intelligent robots. LLMs have demonstrated superior performance in natural language processing and problem-solving abilities, and various attempts are being made in the robotics field to utilize them in path planning. In this study, we propose a framework that enables LLMs to understand map data, plan paths, and provide linguistic explanations for decision-making processes. To achieve this, we enhance the performance of LLMs using incontext learning and fine-tuning methods and verify their performance through various evaluation metrics. Experimental results indicate that LLMs can perform more accurate path planning by providing time-aware path explanations. Future research will design a framework that integrates with ROS (Robot Operating System) and explore methods to condition LLMs to operate on map forms that reflect linguistic context and spatial experience.