IROS15 WS on Perception and Planning for Legged Robot Locomotion in Challenging Domains

Recent advancements in control and actuation allowed legged robots to locomote on very uneven and rough terrain, assuming that either the environment is mostly known or well structured, or proprioception is sufficient for achieving balance. In real-world unstructured environments these assumptions may not hold. Exteroceptive perception is crucial for detecting foothold and handhold affordances in the environment, and generating agile motions accordingly. One of the DARPA Robotics Challenge tasks this year is to locomote on flat, but uneven surfaces. In a real-world disaster scenario, legged robots will need to deal with even rougher terrain under significant uncertainty either for static or dynamically changing environments.

This workshop will provide a platform for researchers from perception and planning in legged robotics to disseminate and exchange ideas, evaluating their advantages and drawbacks. This will include methods for detecting footholds and handholds on uneven and rough surfaces for legged robots including bipeds and quadrupeds, often also using arms. The goal is to show various ways from sensing the environment to finding contacts and planning the body and limb trajectories for achieving agile and robust locomotion. The aim is to foster collaboration among researchers that are working on perception and planning for legged robots to advance the state of the art in robot locomotion.

This full day workshop consists of a mixture of presentations on topics including sensing, perception, planning, and motion generation for various types of legged robots designed to work indoors and outdoors. To stimulate interaction, we also organize a poster session to encourage the participation of young researchers and promote the discussion with the speakers and the audience. Moreover we allocate adequate time for questions and discussion to make the workshop as interactive as possible.

Topics of interest
•    sensing for 3D reconstruction and scene modeling
•    proprioceptive and exteroceptive sensing fusion under uncertainty
•    probabilistic approaches to planning under uncertainty
•    localization and mapping for traversability in static or dynamic environments
•    object detection, segmentation, and categorization for collision and obstacle avoidance
•    collision avoidance and self-collision avoidance
•    environment segmentation and classification
•    motion and path planning for high dimensional environments
•    visual learning for foot placement in rough terrain
•    feature extraction and semantic scene understanding and categorization
•    locomotion and non-gaited locomotion planning
•    contact planning and optimization
•    reactive behaviors and emergency behaviors
•    perception and planning benchmarks

1. Oussama Khatib, Shuyun Chung — Stanford University, USA.
2. Maurice Fallon — University of Edinburgh, UK.
3. Jonas Buchli — ETH, Switzerland.
4. Sven Behnke — University of Bonn, Germany.
5. Nicolas Mansard, Olivier Stasse — LAAS-CNRS, France.
6. Dimitrios Kanoulas — IIT, Italy.
7. Maren Bennewitz, Philipp Karkowski — University of Bonn, Germany.
8. Ioannis Havoutis — IIT, Italy.
9. Katie Byl — University of California, Santa Barbara (UCSB), USA.
10. Katja Mombaur — Heidelberg University, Germany.
11. Oskar von Stryk, A. Stumpf, S. Kohlbrecher — TU Darmstadt, Germany.
12. Marco Hutter — ETH Zurich, Switzerland.