Doctoral Research Seminar on "Playing ball-in-the-cup with model-based and model-free methods on a Kuka robot" by Simon Armleder, ICS/TUM
First, I will talk about model-based and model-free learning algorithms for dynamic and high-speed tasks such as the game ball in the cup/kendama. I implemented this task a while ago on our Kuka robot to start exploring dynamic motion primitives and policy optimization algorithms. Second, I will explain the concept behind the Stack of Task, a framework I am currently implementing on our humanoid robot H1 to program whole-body tasks. Finally, I would like to give an outlook about the problems I would like to tackle during my Doctoral Thesis.
January 17, 2022, 11h online via zoom. iI you would like to participate, please get in touch with the organisers.