Achim J. Lilienthal

CueNav is presented, a video model-based navigation framework combining visual cue guided video planning with an embodiment-specific Inverse-Dynamics Model (IDM) that paves the way toward a generalizable navigation framework for longer-horizon planning and embodiment-aware control.

Confirmation-seeking gazes towards robots increase significantly during dialogue, but robot gaze orientation has little effect on human attention in complex tasks.

Highly expressive multimodal communication outperforms unimodal LED signals in legibility and user intuition, but real-world contexts reveal a drop in effectiveness.