Interaction is fundamentally social, and we cannot stay out of it by “doing nothing” since even inaction is treated as socially accountable. Imagine you ask someone a question and they do not react – it is likely that you will then repeat your question. If they continue to stay silent, you might conclude that they deliberately ignore you (and might confront them). A robot that acts at odds with human expectations of what should happen next is hard to interact and collaborate with. My work sheds light on interactional challenges for robots in real-life settings such as roads, workplaces and homes.
Conversation analysis works with video data to get detailed insights into how humans and machines interact on a moment-by-moment basis. I have collected data in a variety of settings, such as the home, in regular traffic and in surgical operating rooms. For my thesis, I work with Cozmo toy robots and autonomous buses by Navya and EasyMile. I explore how sound can enable smooth interaction and embodied coordination between robots and humans.