This is hardcore science coming from the Institute of Robotics and Mechatronics in Germany, and it’s specifically designed to investigate one very worrisome thing: Can a robot wielding a sharp tool for its job work out that it’s hitting the wrong target–i.e. delicate, fragile human flesh–and stop itself before it accidentally inflicts harm? This is a much more tricky collision avoidance system to test and put into action than the radar parking systems you see on some cars, thanks to the fact the machine has to detect that its cutting tool is not connecting with the right target, and that requires fine sensor powers and fast computing.
Seriously though, if we’re to integrate more and more robots into our daily lives, particularly at the sort of butler/maid level in the home, then this sort of research is absolutely vital.
