Master-Praktikum: Robot Learning from Demonstration for Household Tasks

General-purpose service robots require efficient algorithms for skill acquisition to enable them to perform everyday tasks reliably in unstructured, human-centered environments. A promising approach is the combination of programming by demonstration, intuitive programming guided by prior knowledge and novel control algorithms. This course gives you the chance to implement novel algorithms for service robotics, with a particular focus on contact-rich manipulation tasks found in household environments, such as those encountered in kitchens or bathrooms — for example wiping surfaces, scrubbing sinks and tiles, cleaning mirrors, and handling deformable or slippery objects. Making use of prerecorded demonstration data, you will develop and evaluate approaches that allow robots to generalize cleaning skills to novel objects and surfaces. Algorithms should be implemented in a prototypical fashion on the Kuka iiwa lightweight robot. The course will be conducted in close collaboration with the Interactive Skill Learning group at DLR.

TUMonline entry: https://campus.tum.de/tumonline/ee/ui/ca2/app/desktop/#/slc.tm.cp/student/courses/950974712?$ctx=lang=EN&$scrollTo=toc_overview 

Preliminary Meeting and Registration