IEEE-RAS International Conference on Humanoid Robots (Humanoids 2008)

Workshop
Imitation and Coaching in Humanoid Robots

Monday, December 1st, Daejeon, Korea

Humanoid robotics has dealt with the problem of learning complex humanoid behaviors for a long time. It was soon realized that to overcome problems arising from high dimensional and continuous perception-action spaces, it is necessary to guide the search process and also to develop higher-level representations suitable for faster learning. To achieve these goals, researchers in sensorimotor learning have explored various solutions. Some of the most notable among those are learning by demonstration (or imitation learning) and coaching.

Enabling technologies for imitation learning and coaching such as motion capture, real-time humanoid vision, and computational learning have seen significant progress in recent years. On the other hand, the discovery of mirror systems in humans motivated neuroscience researchers to develop a number of computational models of imitation. These developments should provide new incentives for robotics researchers to develop imitation learning systems using natural interfaces and based on computational theories of human imitation.

This one-day workshop will provide a forum for researchers working in the various areas pertaining to imitation learning and coaching, to interact and exchange ideas

Topics of Interest:

• Movement primitives and imitation;
• Action recognition and understanding;
• Imitation learning in high-dimensional spaces;
• Pruning the search space by coaching;
• Coaching interfaces;
• Task-level imitation;
• Goal-directed imitation;
• Vision for imitation learning;
• Computational models of imitation learning;
• Mirror system and imitation;

Real-time human control of robots for robot skill synthesis
Erhan Oztop, Joshua Hale, Jan Babic, Mitsuo Kawato
ATR, Japan

Incremental learning of full-body human motion primitives for humanoid robots
Dana Kulic
Nakamura-Yamane Laboratory, University of Tokyo, Japan

Transfer of Human Movements to Humanoid Robots
M. Do, D. Gehrig, H. Kühne, P. Azad, P. Pastor, T. Asfour, T. Schultz, A. Wörner, R. Dillmann
University of Karlsruhe, Karlsruhe, Germany

Online Recognition of Daily-Life Movements
D. Gehrig, A. Fischer, H. Kühne, T. Stein, A. Wörner, H. Schwameder and T. Schultz
University of Karlsruhe, Karlsruhe, Germany

Motion Imitation and Recognition using Parametric Hidden Markov Models
Dennis Herzog, Ales Ude, Volker Krüger
Aalborg University Copenhagen, Denmark, Jozef Stefan Institute, Slovenia

Toward coaching a robot by utilizing bioelectrical signals
Kenji Suzuki
Dept. of Intelligent Interaction Technologies, University of Tsukuba, Japan

Learning Through Coaching in Cooperative Side-by-Side Human-Humanoid Interaction
Peter Ford Dominey, Giorgio Metta, Francesco Nori, Lorenzo Natale
CNRS & INSERM U846, France, Italian Institute of Technology, Italy

End of workshop

Prospective participants are required to submit a one page abstract and a three-slide presentation. The presentation should be prepared as three PowerPoint-style slides (Objectives, Approach, Results/
Experiments). Please send your abstract and three-slide presentation directly to the Workshop organizers (asfour@ira.uke.de). The abstracts will be published on the workshop webpage.

Important Dates

• September 20th, 2008 Deadline for submission of one-page abstract and three-slide presentation
• October 5th, 2008 Notification of acceptance
• November 10th, 2008 Deadline for submission of final abstracts
• December 1st, 2008 Workshop

Presentations and abstracts must be submitted in pdf format and not exceed a size of 2Mb.

We warmly thank you for submitting your abstract to the workshop at the Humanoids 2008 and are looking forward to your contribution and an opportunity to create a successful event together with you.

Tamim Asfour
Aleš Ude
Gordon Cheng
Rüdiger Dillmann

For any further information, please contact the organizers Tamim Asfour (asfour at ira.uka.de) or Aleš Ude (ales.ude@ijs.si)