This research integrates large language models with reinforcement learning to create robots that navigate social spaces through natural conversation.

• Introduces HSAC-LLM framework combining deep reinforcement learning with language models
• Enables bidirectional conversations between humans and robots in shared spaces
• Allows robots to understand social context, explain decisions, and respond to human queries
• Demonstrates improved navigation safety while maintaining natural human-robot interactions

This advancement matters because it transforms robots from mere obstacle-avoiding machines into socially aware entities that can communicate intentions and understand human feedback—critical for the next generation of service robots in public spaces.

Socially-Aware Robot Navigation Enhanced by Bidirectional Natural Language Conversations Using Large Language Models