Learning (in) Task and Motion Planning

RSS 2020 Workshop

July 12, 2020, Oregon State University at Corvallis, Oregon, USA


Task and Motion Planning (TAMP) frameworks show remarkable capabilities in scaling to long action sequences, many objects and a variety of tasks. However, TAMP usually assumes perfect knowledge, relies on simplified (kinematic) models of the world, requires long computation time and most of the time yields open-loop motion plans, all of which limit the robust and practical applicability of TAMP in the real world.

On the other end of the spectrum, reinforcement learning (RL) techniques have demonstrated, also in real world experiments, the ability to solve manipulation problems with complex contact interactions in a robust and closed-loop fashion. The disadvantage of most of these approaches is that they work for a single goal only, require huge amounts of trials and have trouble showing the same long-term sequential planning behaviors of classical TAMP frameworks.

The goal of this workshop is to investigate if and how learning can address the challenges imposed by TAMP problems to develop (novel) methods that achieve both the generality of TAMP approaches and the complex interaction capabilities of RL policies.

To discuss this, we are trying to bring together experts from the fields of

  • TAMP
  • Meta/multi-goal reinforcement learning
  • Feedback motion planning (with special interest in contact/force based planning)
  • Perception for planning
as well as experts that examine the boundaries between these.

This workshop continues past RSS workshops on Task and Motion Planning (2016, 2017, 2018, 2019) with a focus on learning this year.

Invited Speakers


09:00 - 09:10Introduction
09:10 - 09:50Dieter Fox
09:50 - 10:30Russ Tedrake
10:30 - 10:40Discussion 1
10:40 - 11:00Coffee break
11:00 - 11:40Lydia Tapia
11:40 - 12:20Tomas Lozano-Perez
12:20 - 12:30Discussion 2
12:30 - 02:00Workshop lunch
02:00 - 02:40Sergey Levine
02:40 - 03:20Jeannette Bohg
03:20 - 03:30Discussion 3
03:30 - 04:00Poster spotlight presentations
04:00 - 04:30Poster session & coffee
04:30 - 05:10Georg von Wichert
05:10 - 05:50Weiwei Wan
05:50 - 06:00Discussion 4 and summary

Call for Contributions

We solicit 2-3 page extended abstracts using the standard RSS template. References do not count to the page limit.

Topics of interest (but not limited to) are

  • Task and Motion Planning
  • Learned heuristics for TAMP
  • Learning TAMP
  • Meta/multi-goal reinforcement learning
  • Perception for planning
  • Feedback motion planning
  • Real world execution of TAMP
  • TAMP with force interactions/physics
  • Robot manipulation

Important dates:

Submission deadline: April 9, 2020. Anywhere on Earth
Acceptance notification: April 16, 2020

Note that due to possible visa processing times, this submission deadline is pretty early.

Submissions should be sent directly to Danny Driess as a PDF-file. Additional video attachments are welcome.

We seek original research, late breaking results that still need discussion or work that discusses the workshop topic (with empirical data or theoretical foundation). An overlap with submitted/accepted papers is acceptable, if they have not been presented before.

Accepted contributions (and optional video attachments) will be published on this website.