Robots to learn to sense when something goes wrong

A new research project from Aarhus University and the University of Southern Denmark (SDU) combines adaptive AI with soft robotic skin. The aim is to build robots that can detect physical changes and respond to them.

That is all the source text gives. It does not say who leads the project, how it is funded, how long it will run, or what kind of robots or sensors will be used. It also does not describe what counts as "something going wrong", whether that means a collision, a slipping grip, a damaged part or another change, nor how the robots would react once they notice it.

The idea behind the pairing is that soft skin gives a robot a way to feel its physical surroundings, while adaptive AI lets it adjust its behaviour when conditions change instead of following fixed instructions. The source does not report any results, prototypes or test settings, so the work should be read as a newly described research effort rather than a finished technology.

The project is relevant to Danish AI and robotics research because it brings together two of the country's major universities around a shared problem: making robots more aware of their physical state and better able to cope with the unexpected. The item was first published by TechSavvy under the Danish title "Robotter skal lære at mærke, når noget går galt."