This research introduces a dual-core framework that leverages large language models to enhance navigation systems, enabling them to handle non-predefined scenarios that traditional systems struggle with.

• Employs ReAct-based prompt engineering to break complex navigation tasks into manageable sub-tasks
• Creates an adaptive system that can autonomously invoke external tools as needed
• Delivers more robust navigation support across unfamiliar scenarios
• Enhances safety and reliability in engineering applications

This innovation represents a significant advancement in navigation engineering by creating more flexible, responsive systems that can intelligently adapt to unexpected conditions and requirements.

Navigation-GPT: A Robust and Adaptive Framework Utilizing Large Language Models for Navigation Applications