A Long Horizon Planning Framework for Manipulating Rigid Pointcloud Objects

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 42
  • Download : 0
DC FieldValueLanguage
dc.contributor.authorSimeonov, Anthonyko
dc.contributor.authorDu, Yilunko
dc.contributor.authorKim, Beomjoonko
dc.contributor.authorHogan, Francoisko
dc.contributor.authorTenenbaum, Joshuako
dc.contributor.authorAgrawal, Pulkitko
dc.contributor.authorRodriguez, Albertoko
dc.date.accessioned2021-06-14T00:30:18Z-
dc.date.available2021-06-14T00:30:18Z-
dc.date.created2021-06-14-
dc.date.created2021-06-14-
dc.date.issued2020-11-16-
dc.identifier.citation4th Conference on Robot Learning (CoRL 2020)-
dc.identifier.urihttp://hdl.handle.net/10203/285812-
dc.description.abstractWe present a framework for solving long-horizon planning problems involving manipulation of rigid objects that operates directly from a point-cloud observation, i.e. without prior object models. Our method plans in the space of object subgoals and frees the planner from reasoning about robot-object interaction dynamics by relying on a set of generalizable manipulation primitives. We show that for rigid bodies, this abstraction can be realized using low-level manipulation skills that maintain sticking contact with the object and represent subgoals as 3D transformations. To enable generalization to unseen objects and improve planning performance, we propose a novel way of representing subgoals for rigid-body manipulation and a graph-attention based neural network architecture for processing point-cloud inputs. We experimentally validate these choices using simulated and real-world experiments on the YuMi robot. Results demonstrate that our method can successfully manipulate new objects into target configurations requiring long-term planning. Overall, our framework realizes the best of the worlds of task-and-motion planning (TAMP) and learning-based approaches.-
dc.languageEnglish-
dc.publisherMassachusetts Institute of Technology-
dc.titleA Long Horizon Planning Framework for Manipulating Rigid Pointcloud Objects-
dc.typeConference-
dc.type.rimsCONF-
dc.citation.publicationname4th Conference on Robot Learning (CoRL 2020)-
dc.identifier.conferencecountryUS-
dc.identifier.conferencelocationVirtual-
dc.contributor.localauthorKim, Beomjoon-
dc.contributor.nonIdAuthorSimeonov, Anthony-
dc.contributor.nonIdAuthorDu, Yilun-
dc.contributor.nonIdAuthorHogan, Francois-
dc.contributor.nonIdAuthorTenenbaum, Joshua-
dc.contributor.nonIdAuthorAgrawal, Pulkit-
dc.contributor.nonIdAuthorRodriguez, Alberto-
Appears in Collection
RIMS Conference Papers
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0