Please use this identifier to cite or link to this item: https://doi.org/10.15480/882.2954
DC FieldValueLanguage
dc.contributor.authorAgboh, Wisdom-
dc.contributor.authorGrainger, Oliver-
dc.contributor.authorRuprecht, Daniel-
dc.contributor.authorDogar, Mehmet R,-
dc.date.accessioned2020-10-01T11:37:00Z-
dc.date.available2020-10-01T11:37:00Z-
dc.date.issued2020-09-23-
dc.identifier.citationComputing and Visualization in Science 1-4 (23): 8 (2020-09-23)de_DE
dc.identifier.issn1433-0369de_DE
dc.identifier.urihttp://hdl.handle.net/11420/7444-
dc.description.abstractA key component of many robotics model-based planning and control algorithms is physics predictions, that is, forecasting a sequence of states given an initial state and a sequence of controls. This process is slow and a major computational bottleneck for robotics planning algorithms. Parallel-in-time integration methods can help to leverage parallel computing to accelerate physics predictions and thus planning. The Parareal algorithm iterates between a coarse serial integrator and a fine parallel integrator. A key challenge is to devise a coarse model that is computationally cheap but accurate enough for Parareal to converge quickly. Here, we investigate the use of a deep neural network physics model as a coarse model for Parareal in the context of robotic manipulation. In simulated experiments using the physics engine Mujoco as fine propagator we show that the learned coarse model leads to faster Parareal convergence than a coarse physics-based model. We further show that the learned coarse model allows to apply Parareal to scenarios with multiple objects, where the physics-based coarse model is not applicable. Finally, we conduct experiments on a real robot and show that Parareal predictions are close to real-world physics predictions for robotic pushing of multiple objects. Videos are at https://youtu.be/wCh2o1rf-gA.en
dc.language.isoende_DE
dc.publisherSpringerde_DE
dc.relation.ispartofComputing and visualization in sciencede_DE
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/de_DE
dc.subjectLearningde_DE
dc.subjectManipulationde_DE
dc.subjectModel-predictive controlde_DE
dc.subjectNeural networkde_DE
dc.subjectParallel-in-timede_DE
dc.subjectPararealde_DE
dc.subjectPlanningde_DE
dc.subjectRoboticsde_DE
dc.subjectComputer Science - Roboticsde_DE
dc.subjectComputer Science - Roboticsde_DE
dc.subjectComputer Science - Learningde_DE
dc.subject.ddc510: Mathematikde_DE
dc.titleParareal with a learned coarse model for robotic manipulationde_DE
dc.typeArticlede_DE
dc.identifier.doi10.15480/882.2954-
dc.type.diniarticle-
dcterms.DCMITypeText-
tuhh.identifier.urnurn:nbn:de:gbv:830-882.0107213-
tuhh.oai.showtruede_DE
tuhh.abstract.englishA key component of many robotics model-based planning and control algorithms is physics predictions, that is, forecasting a sequence of states given an initial state and a sequence of controls. This process is slow and a major computational bottleneck for robotics planning algorithms. Parallel-in-time integration methods can help to leverage parallel computing to accelerate physics predictions and thus planning. The Parareal algorithm iterates between a coarse serial integrator and a fine parallel integrator. A key challenge is to devise a coarse model that is computationally cheap but accurate enough for Parareal to converge quickly. Here, we investigate the use of a deep neural network physics model as a coarse model for Parareal in the context of robotic manipulation. In simulated experiments using the physics engine Mujoco as fine propagator we show that the learned coarse model leads to faster Parareal convergence than a coarse physics-based model. We further show that the learned coarse model allows to apply Parareal to scenarios with multiple objects, where the physics-based coarse model is not applicable. Finally, we conduct experiments on a real robot and show that Parareal predictions are close to real-world physics predictions for robotic pushing of multiple objects. Videos are at https://youtu.be/wCh2o1rf-gA.de_DE
tuhh.publisher.doi10.1007/s00791-020-00327-0-
tuhh.publication.instituteMathematik E-10de_DE
tuhh.identifier.doi10.15480/882.2954-
tuhh.type.opus(wissenschaftlicher) Artikel-
dc.type.driverarticle-
dc.type.casraiJournal Article-
tuhh.container.issue1-4de_DE
tuhh.container.volume23de_DE
dc.rights.nationallicensefalsede_DE
dc.identifier.arxiv1912.05958v2de_DE
dc.identifier.scopus2-s2.0-85091274340de_DE
tuhh.container.articlenumber8de_DE
local.status.inpressfalsede_DE
local.type.versionpublishedVersionde_DE
datacite.resourceTypeJournal Article-
datacite.resourceTypeGeneralText-
item.grantfulltextopen-
item.openairecristypehttp://purl.org/coar/resource_type/c_6501-
item.creatorGNDAgboh, Wisdom-
item.creatorGNDGrainger, Oliver-
item.creatorGNDRuprecht, Daniel-
item.creatorGNDDogar, Mehmet R,-
item.openairetypeArticle-
item.fulltextWith Fulltext-
item.cerifentitytypePublications-
item.creatorOrcidAgboh, Wisdom-
item.creatorOrcidGrainger, Oliver-
item.creatorOrcidRuprecht, Daniel-
item.creatorOrcidDogar, Mehmet R,-
item.languageiso639-1en-
item.mappedtypeArticle-
crisitem.author.deptMathematik E-10-
crisitem.author.orcid0000-0002-0242-0215-
crisitem.author.orcid0000-0003-1904-2473-
crisitem.author.orcid0000-0002-6896-5461-
crisitem.author.parentorgStudiendekanat Elektrotechnik, Informatik und Mathematik-
Appears in Collections:Publications with fulltext
Files in This Item:
File Description SizeFormat
Agboh2020_Article_PararealWithALearnedCoarseMode.pdfVerlagsversion1,48 MBAdobe PDFView/Open
Thumbnail
Show simple item record

Page view(s)

191
Last Week
1
Last month
3
checked on Aug 15, 2022

Download(s)

106
checked on Aug 15, 2022

SCOPUSTM   
Citations

2
Last Week
0
Last month
0
checked on Jul 11, 2022

Google ScholarTM

Check

Note about this record

Cite this record

Export

This item is licensed under a Creative Commons License Creative Commons