Next-generation robots move the places of their activity from the industry to the places that are closely related to people, such as welfare and nursing care. Their control including physical human-robot interaction (pHRI) is so complicated as to be too difficult for pure model-based control. To resolve such complexity, machine learning techniques like deep learning are powerfully exploited, while end-to-end learning is tough work in the real world. In this talk, I introduce case studies of the machine-learning-used optimization of model-based control. Machine learning techniques are utilized to optimize the parameters in the given model-based control, thereby achieving appropriate control with less burden on the user in pHRI.
Taisuke Kobayashi received his B.E., M.E., and Ph.D. degrees from Nagoya University, Japan, in 2012, 2014, and 2016, respectively. He is currently an Assistant Professor of Nara Institute of Science and Technology (NAIST), Japan. From May 2018 to March 2019, he is working at ICS, TUM as a member of Program for Advancing Strategic International Networks to Accelerate the Circulation of Talented Researchers.
His research interests include the locomotion control by intelligent systems and autonomous robotics with reinforcement learning.
Host: Alexander Plopski