Research

Research with Everyday Relevance

We are exploring the co-construction between humans and robots using an everyday example: joint activities in the kitchen. In the future, AI-based robots could help us with cooking or setting the table. This would be of enormous importance, especially for older and physically impaired people.
However, what sounds like simple tasks at first, often poses major challenges for today's robots: Stovetops are hot and glasses are fragile - what we humans learn as children, robots also have to learn. We humans intuitively understand that water flows faster than cake batter - robots do not yet have intuition.
And it takes a lot more than that to be able to apply this knowledge correctly: some people like butter with their jam sandwich, others don't. And a candlelit dinner needs to be prepared differently to an everyday breakfast. For robots to be able to support us flexibly in such scenarios, they not only need factual knowledge about their environment, but also the ability to develop an intelligent understanding of their environment, their tasks. and the consequences of their actions together with humans.