DC Field | Value | Language |
---|---|---|
dc.contributor.author | Simeonov, Anthony | ko |
dc.contributor.author | Du, Yilun | ko |
dc.contributor.author | Kim, Beomjoon | ko |
dc.contributor.author | Hogan, Francois | ko |
dc.contributor.author | Tenenbaum, Joshua | ko |
dc.contributor.author | Agrawal, Pulkit | ko |
dc.contributor.author | Rodriguez, Alberto | ko |
dc.date.accessioned | 2021-06-14T00:30:18Z | - |
dc.date.available | 2021-06-14T00:30:18Z | - |
dc.date.created | 2021-06-14 | - |
dc.date.created | 2021-06-14 | - |
dc.date.issued | 2020-11-16 | - |
dc.identifier.citation | 4th Conference on Robot Learning (CoRL 2020) | - |
dc.identifier.uri | http://hdl.handle.net/10203/285812 | - |
dc.description.abstract | We 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.language | English | - |
dc.publisher | Massachusetts Institute of Technology | - |
dc.title | A Long Horizon Planning Framework for Manipulating Rigid Pointcloud Objects | - |
dc.type | Conference | - |
dc.type.rims | CONF | - |
dc.citation.publicationname | 4th Conference on Robot Learning (CoRL 2020) | - |
dc.identifier.conferencecountry | US | - |
dc.identifier.conferencelocation | Virtual | - |
dc.contributor.localauthor | Kim, Beomjoon | - |
dc.contributor.nonIdAuthor | Simeonov, Anthony | - |
dc.contributor.nonIdAuthor | Du, Yilun | - |
dc.contributor.nonIdAuthor | Hogan, Francois | - |
dc.contributor.nonIdAuthor | Tenenbaum, Joshua | - |
dc.contributor.nonIdAuthor | Agrawal, Pulkit | - |
dc.contributor.nonIdAuthor | Rodriguez, Alberto | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.