RAIDER integrates Large Language Models with grounded tools to help robots address action-related issues in dynamic human environments.

• Combines LLMs with real-world constraints for improved issue detection
• Enables robots to provide clear explanations of failures
• Implements adaptive recovery strategies for robotic systems
• Designed for dynamic environments where traditional methods fall short

This research advances robotic reliability by creating more adaptable systems that can function effectively alongside humans, detect problems earlier, and recover more efficiently from failures.

RAIDER: Tool-Equipped Large Language Model Agent for Robotic Action Issue Detection, Explanation and Recovery