Bionic Learning for the Control of the Bionic Handling Assistant (BHA)

Download Factsheet PDF

Platform:
The „Bionic Handling Assistant“ is a new robotic platform manufactured by Festo. Its pneumatic actuation and flexible, lightweight material permit a safe physical interaction between the robot and humans.

Challenges for Research:

• This bionic „elephant trunk“ challenges standard approaches, since its behavior can only be described approximately.
• Machine learning approaches allow to deal with little analytical knowledge and offer fast and accurate control.
• New research results for such learning methods thereby allow to make strict guarantees about the behavior
• Grasping for objects demands for a highly coordinated use of the many actuators on the robot.

   

  Our Research Goal:
  The scientists at the research institut CoR-Lab show that new learning methods, inspired by research on infants' early grasping attempts, allow for an efficient bootstrapping of such skills, even though the set of possible movements is too large to be ever explored exhaustively.

  BHA Automatica Exhibit:
  We show live at Automatica 2012 how this learning proceeds on the robot, and how in paves the way for the practical application of the Bionic Handling Assistant.

  Contact:
  Research Institute for Cognition and Robotics (CoR-Lab), Bielefeld University
  Dipl.-Inform. Matthias Rolf:
  mrolf@cor-lab.uni-bielefeld.de

  Prof. Dr. Jochen J. Steil:
  jsteil@cor-lab.uni-bielefeld.de
  www.cor-lab.de