[robotics-worldwide] [meetings] IROS Workshop on Compliant Manipulation: Challenges in Learning and Control

Klas Kronander klas.kronander at epfl.ch
Mon Sep 8 09:25:34 PDT 2014

Call for participation
IROS 2014 workshop on Compliant Manipulation: Challenges in Learning and
Chicago, September 18 2014
website: http://lasa.epfl.ch/workshopIROS14/
contact: klas.kronander at epfl.ch

Accepted contributions and program are now online at 

Robust and versatile compliant manipulation skills are a necessity for
robots interacting with the real world. Despite significant progress in
design of passively compliant mechanisms and active compliance control
algorithms, today's best robots are still far behind humans in terms of
manipulation performance and versatility. While using compliant
joints/control as opposed to traditional position-control offers the
possibility to perform more complex tasks, it also makes the task
specification a much more challenging problem.

Learning is often advocated as a means of dealing with the complexity of
control and planning tasks in robotics. Various forms of learning such
as Learning from Demonstration and Reinforcement Learning have been
integrated in robotics for a long time with great success in many
applications. This workshop will focus on how different forms of
learning can be used to advance the state of the art in compliant

The goal of this full day workshop is to bring together experts in the
yet largely disconnected fields of control/design and robot task
learning, and to foster increased knowledge exchange and collaboration
between these communities. Additionally, the workshop will pinpoint open
challenges in compliant manipulation. The speakers include experts from
learning and control in robots and humans. In addition to invited talks,
contributions will be presented in interactive poster sessions.

Neville Hogan, MIT
Jonas Buchli, ETH Zurich
Yoshihiko Nakamura, University of Tokyo
Etienne Burdet, Imperial College London
Sylvain Calinon, IDIAP research institute
Matthew Howard, Kings College

-  Novel architectures for control of compliant manipulation
-  Passively compliant systems
-  Impedance and interaction control
-  Task transfer: Learning by demonstration and interaction
-  Task transfer: Reinforcement learning
-  Novel task representations suitable for learning and control
-  Adaptive control for manipulation
-  Biologically inspired control algorithms for manipulation

Klas Kronander, EPFL
Aude Billard, EPFL
Etienne Burdet, Imperial College London
Jonas Buchli,  ETHZ

This workshop is supported by the Swiss National Centre of Competence in
Research Robotics, www.nccr-robotics.ch.

More information about the robotics-worldwide mailing list