Summary
Techniques that enable learning in simulation taking into account uncertainties will be described. The aim is to compute position and force control specifications which afterwards only require a relative simple adjustment on the real platform