Oya Celiktutan

Towards Socially Adaptive Robots: Behaviour Modelling, Interaction, and Evaluation

Talk abstract:
As robots move into human-centred environments such as healthcare, their ability to behave in socially meaningful ways becomes critical. Beyond perception, this requires adapting to social norms, using cues such as gaze and gesture, recognising emotions, and explaining actions in ways that are intuitive and acceptable to users. In this talk, I will present examples from our research on how these capabilities can be implemented in practice, focusing on robots that can demonstrate social awareness, comply with social norms, understand and imitate non-verbal behaviour, and communicate their intentions. I will show how these behaviours shape user perception and engagement, particularly in real-world settings. I will conclude by outlining the key challenges in building robots that can adapt socially in complex environments.