ICRA 2011 Paper Abstract


Paper TuA1-InteracInterac.6

Do, Martin (Karlsruhe Institute of Technology (KIT)), Asfour, Tamim (Karlsruhe Institute of Technology (KIT)), Dillmann, Rüdiger (KIT Karlsruher Institut für Technologie)

Towards a Unifying Grasp Representation for Imitation Learning on Humanoid Robots

Scheduled for presentation during the Poster Sessions "Interactive Session I: Robotic Technology" (TuA1-InteracInterac), Tuesday, May 10, 2011, 08:20−09:35, Hall

2011 IEEE International Conference on Robotics and Automation, May 9-13, 2011, Shanghai International Conference Center, Shanghai, China

This information is tentative and subject to change. Compiled on April 2, 2020

Keywords Grasping, Humanoid Robots, Learning and Adaptive Systems


In this paper, we present a grasp representation in task space exploiting information on finger tips. The proposed representation provides a suitable basis for imitation learning methods since parameters representing a specific grasp type can be estimated merely from the images originating from the robot’s onboard stereo camera setup. For this purpose, a motion tracking algorithm is introduced and an estimation procedure is outlined in this work. Inspired by neuroscientific studies, synergies on task space level are established to accommodate for correlations within the finger movements. Introduced into the representation, it allows the interpretation as a motor control policy with a low-dimensional control variable, which merges the different grasp stages, preshape, reach, and enclose as a kinematic unit. Experiments are conducted showing that the smooth trajectories emerging from the system feature humanlike characteristics.



Technical Content © IEEE Robotics & Automation Society

This site is protected by copyright and trademark laws under US and International law.
All rights reserved. © 2002-2020 PaperCept, Inc.
Page generated 2020-04-02  13:07:02 PST  Terms of use