In order to learn optimized motion sequences, robots must be able to explore their environment autonomously.
A method developed at IPR now ensures that neither the robots nor their surroundings are damaged during this exploration process.