Workshop : Offline and online optimization for humanoid robots

  • Eiichi Yoshida, AIST Tsukuba, Japan
  • Katja Mombaur, IWR, University of Heidelberg, Germany
  • Yuval Tassa, Tom Erez, University of Washington, USA,

The workshop is jointly organized by the two IEEE RAS Technical Committees on Humanoid Robots and on Model-based Optimization for Robotics.

Time and Place:
  • November 7, 2013, 8:50 am – 5 pm, detailed program see below
  • Tokyo Big Sight, Tokyo, Japan, exact room t.b.a.
  • More information about the conference at
Objectives of Workshop:

Humanoid robots are very complicated dynamical systems for which the generation of motions is a difficult task, since the number of parameters to tune for a motion is very high. The target application areas for humanoid robots today and in the future require them to move in a variety of complex terrains and to automatically generate and control a wide range of motions in order to be more flexible and adaptive to changing environments. Optimization or optimal control offers an interesting way to generate motions automatically based on elementary principles described by cost functions and constraints. Advances in computing equipment as well as recent developments in optimization algorithms and in particular real-time optimization make a wider application of algorithmic optimization a realistic option even for real-time control in complex robotic applications in the near future.

The objective of this workshop is to gather researchers from humanoid robotics and from optimization in order to present latest algorithmic developments in offline and online optimization for humanoid robotics as well as current applications and future challenging application areas still waiting for appropriate methods. We also intend to include researchers from adjacent research fields such as computer graphics in order to discuss the use of optimization approaches for virtual anthropomorphic systems developed in these areas for real-life robotic systems. The methods discussed may also be interesting for researchers in other fields of dynamic legged robotics, such as quadruped robots.

  • Developing good dynamical models of the robot and its environment for optimization
  • Optimization based generation of humanoid robot trajectories
  • Improving the behavior style of humanoid robots by optimization, e.g. induce “natural” behavior or emotional body language
  • Optimization based design of humanoid robots
  • Online motion control using real-time model-based optimization and model predictive control / receding horizon control
  • Development of appropriate dynamical models for offline and online optimization
  • Learning /improving models during optimization
  • Inverse optimal control techniques for the identification of objective functions from human motions for the use in humanoid robots
  • Robust optimal control and refinement of optimal controls based on actual experience
  • Parameter estimation and optimum experimental design for humanoid robots
Confirmed workshop speakers
  • Ken’ichiro Nagasaka, Sony, Japan
  • Michiel van de Panne, UBC, Vancouver, Canada
  • Jean-Paul Laumond, LAAS-CNRS, Toulouse, France
  • Francesco Nori IIT, Genova, Italy
  • Russ Tedrake, MIT, Boston, USA
  • Adrien Escande, CNRS-AIST JRL, Tsukuba, Japan
  • Tamim Asfour, KIT, Germany
  • Dinesh Manosha, University of North Carolina, USA
  • Eiichi Yoshida, AIST, Tsukuba, Japan
  • Katja Mombaur and Henning Koch, University of Heidelberg, Germany
  • Tom Erez and Yuval Tassa, University of Washington, Seattle, USA

Talk abstracts are available here

Intended Audience:

This workshop aims at gathering audience (researchers, students and industry) from the fields of

  • humanoid robots
  • optimization / applied mathematics / scientific computing
  • computer graphics
  • computer science
  • other fields of legged robotics

interested in the generation and control of anthropomorphic motions and other systems of similar complexity.

The workshop is organized as part of the European project KoroiBot