In this thesis, you will investigate how multiple mobile robots can navigate safely and naturally in environments shared with humans. Human navigation is shaped by implicit social conventions: people slow down near uncertain motion, avoid cutting too closely in front of others, yield in narrow spaces, and adapt their trajectories based on the behavior and intent of those around them. These rules are rarely stated explicitly, yet they strongly influence how we move in crowds.

The goal of this thesis is to explore how Large Language Models (LLMs) can help ground robots interpret and apply such social navigation rules in a human-aware manner. The student will design navigation strategies that incorporate common-sense reasoning for socially acceptable robot motion, such as adapting speed, choosing passing direction, yielding in shared spaces, and reacting to human intent. These behaviors will be integrated into a multi-robot navigation framework and evaluated in simulation and, optionally, on real ground rovers.

Possible tasks include:

• Designing socially aware navigation rules for multi-robot systems in human environments
• Using LLMs to generate context-aware navigation decisions from human-centered scene descriptions
• Integrating human-aware behavior into local planning or multi-robot coordination
• Evaluating robot behavior in terms of safety, efficiency, and social acceptability
• Testing in simulation and optionally on real outdoor ground robots

If you are interested in this topic, please send an email to both Prof. Groß and Ecem Isildar including a brief motivation letter, your CV, and your current transcript of records.

Core data