Optimal Robot Motion Planning Workshop

An ICRA 2015 Workshop

A half-day workshop, which will take place Saturday, May 30, 2015 in conjunction with the IEEE International Conference on Robotics and Automation (ICRA) 2015.

Description

As modern robots address real-world problems in dynamic, unstructured, and open environments, a key requirement is the ability to quickly plan highly efficient motions. There has been recent renewed interest and a growing variety of methods for motion planning, which vary in terms of their optimality properties. In particular:

  • Search-based methods typically operate over a discretization of a continuous configuration space. They provide optimal paths over the discretization but they face challenges in scaling to higher-dimensions. Recent developments, however, have shown promise in solving manipulation challenges.
  • Sampling-based planners have traditionally focused on quickly finding a feasible but suboptimal collision-free path in geometrically complex scenes for high-dimensional systems. Recently the focus has shifted towards also achieving asymptotic optimality.
  • Optimization-based solutions solve motion-planning queries with high quality paths really fast. While the methodology does not guarantee global optimality, recent techniques exhibit good experimental performance and desirable tradeoffs in terms of path quality and computational overhead.

This workshop aims to bring together researchers from these different areas to establish the state-of-the-art and identify new opportunities at the intersection of existing solutions. The objective is methodologies with stronger guarantees, improved experimental performance and which address increasingly more complex challenges for real-world robots.

Organizers

Chris Dellin
Doctoral Candidate
The Robotics Institute, Carnegie Mellon University
Email: cdellin\AT\cs.cmu.edu
Andrew Dobson
Doctoral Candidate
Computer Science, Rutgers University
Email: chuples.dobson\AT\cs.rutgers.edu
Siddhartha Srinivasa
Finmeccanica Associate Professor in Computer Science
The Robotics Institute, Carnegie Mellon University
Email: siddh\AT\cs.cmu.edu
Kostas Bekris
Assistant Professor in Computer Science
Rutgers, the State University in New Jersey
Email: kostas.bekris\AT\cs.rutgers.edu

Structure

The workshop will include talks from invited speakers that are active in the different subareas of motion planning, as well as contributed talks from speakers who will submit abstracts for interactive sessions. The interactive session presenters will have the opportunity to introduce the topic of their work during short spotlight presentations.

Schedule:

[8:15 – 8:35] Opening Remarks

Session A: On Experimental Performance and Real-World Challenges
[8:35 – 9:00] Talk 1: Ron Alterovitz - Asymptotically Optimal Motion Planning for Medical and Assistive Robots
[9:00 – 9:25] Talk 2: Kris Hauser - Performance Bottlenecks for Optimal Motion Planning
[9:25 – 9:50] Talk 3: Dmitry Berenson - Local and Global Methods for Deformable Object Manipulation

Interactive Session
[9:50 – 10:00] Spotlight presentations
[10:00 – 10:30] Interactive session - coffee break

Session B: On Alternative Foundational Methods and Optimality Guarantees
[10:30 – 10:55] Talk 4: Dan Halperin - Compact Roadmaps for Motion Planning: Efficient Summary of High-Quality Paths
[10:55 – 11:20] Talk 5: Hanna Kurniawati - Exploiting Structures for Approximately Optimal Planning in Uncertain Domains
[11:20 – 11:45] Talk 6: Marco Pavone - Deterministic Sampling-Based Motion Planning: Optimality, Complexity, and Performance
[11:45 – 12:10] Talk 7: Max Likhachev - Search-based Planning with Multiple Inadmissible Heuristics

[12:10 - 12:30] Panel Discussion
- Dan Halperin, David Hsu, Steve LaValle, Katsu Yamane

Additional Program Committee Members

Invited Speakers

Panelists

The workshop culminates in a Panel Discussion, which will be lead by a group of top researchers in the field.

  • Dan Halperin, Tel Aviv University, Israel
  • David Hsu, National University of Singapore, Singapore
  • Steve LaValle, University of Illinois at Urbana-Champaign, USA
  • Katsu Yamane, Disney Research and Carnegie Mellon University (CMU), USA

Interactive Session Presentations


Support from IEEE RAS Technical Committees

We have received the support of the following IEEE RAS Technical Committees: