The execution of motor tasks on a mechanical systems is one of the most important topics in motor skill learning. In here, I am interested in both traditional, analytical and novel policy learning approaches. My focus lies on finding architectures which are general, can be used for various tasks and scale in the domain of interesting control systems such as humanoid robots. Among the topics, I am interested in.

Gauss-Udwadia Framework: See here.
Learning Operational Space Control: See here.
Learning for Locomotion: Check out our recent workshop here.

This page is still a stub. It will be updated shortly.

  

zum Seitenanfang