Exploratory Grasping: Asymptotically Optimal Algorithms for Grasping Challenging Polyhedral Objects

There has been significant recent work on data-driven algorithms for learning general-purpose grasping policies. However, these policies can consistently fail to grasp challenging objects which are significantly out of the distribution of objects in the training data or which have very few high qual...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2020-11
Hauptverfasser: Danielczuk, Michael, Balakrishna, Ashwin, Brown, Daniel S, Devgon, Shivin, Goldberg, Ken
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title arXiv.org
container_volume
creator Danielczuk, Michael
Balakrishna, Ashwin
Brown, Daniel S
Devgon, Shivin
Goldberg, Ken
description There has been significant recent work on data-driven algorithms for learning general-purpose grasping policies. However, these policies can consistently fail to grasp challenging objects which are significantly out of the distribution of objects in the training data or which have very few high quality grasps. Motivated by such objects, we propose a novel problem setting, Exploratory Grasping, for efficiently discovering reliable grasps on an unknown polyhedral object via sequential grasping, releasing, and toppling. We formalize Exploratory Grasping as a Markov Decision Process, study the theoretical complexity of Exploratory Grasping in the context of reinforcement learning and present an efficient bandit-style algorithm, Bandits for Online Rapid Grasp Exploration Strategy (BORGES), which leverages the structure of the problem to efficiently discover high performing grasps for each object stable pose. BORGES can be used to complement any general-purpose grasping algorithm with any grasp modality (parallel-jaw, suction, multi-fingered, etc) to learn policies for objects in which they exhibit persistent failures. Simulation experiments suggest that BORGES can significantly outperform both general-purpose grasping pipelines and two other online learning algorithms and achieves performance within 5% of the optimal policy within 1000 and 8000 timesteps on average across 46 challenging objects from the Dex-Net adversarial and EGAD! object datasets, respectively. Initial physical experiments suggest that BORGES can improve grasp success rate by 45% over a Dex-Net baseline with just 200 grasp attempts in the real world. See https://tinyurl.com/exp-grasping for supplementary material and videos.
format Article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2460192674</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2460192674</sourcerecordid><originalsourceid>FETCH-proquest_journals_24601926743</originalsourceid><addsrcrecordid>eNqNzk8LgjAABfARBEn5HQadhbmp_bmJWN3s0F2WTZ1Mt7YJ7du3IDp3eg_e7_AWIMCExNE-wXgFQmMGhBDOdjhNSQCa8qWE1NRK7eBZU6P41B1hbtyorLS8oUI4WCnLRypgLjqpue1HA1upfx4WvWds6j79KoXr2UN7Xt0H1lizAcuWCsPCb67B9lTeikuktHzOzNh6kLOe_FTjJEPxwb9LyH_qDXBrRwM</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2460192674</pqid></control><display><type>article</type><title>Exploratory Grasping: Asymptotically Optimal Algorithms for Grasping Challenging Polyhedral Objects</title><source>Free E- Journals</source><creator>Danielczuk, Michael ; Balakrishna, Ashwin ; Brown, Daniel S ; Devgon, Shivin ; Goldberg, Ken</creator><creatorcontrib>Danielczuk, Michael ; Balakrishna, Ashwin ; Brown, Daniel S ; Devgon, Shivin ; Goldberg, Ken</creatorcontrib><description>There has been significant recent work on data-driven algorithms for learning general-purpose grasping policies. However, these policies can consistently fail to grasp challenging objects which are significantly out of the distribution of objects in the training data or which have very few high quality grasps. Motivated by such objects, we propose a novel problem setting, Exploratory Grasping, for efficiently discovering reliable grasps on an unknown polyhedral object via sequential grasping, releasing, and toppling. We formalize Exploratory Grasping as a Markov Decision Process, study the theoretical complexity of Exploratory Grasping in the context of reinforcement learning and present an efficient bandit-style algorithm, Bandits for Online Rapid Grasp Exploration Strategy (BORGES), which leverages the structure of the problem to efficiently discover high performing grasps for each object stable pose. BORGES can be used to complement any general-purpose grasping algorithm with any grasp modality (parallel-jaw, suction, multi-fingered, etc) to learn policies for objects in which they exhibit persistent failures. Simulation experiments suggest that BORGES can significantly outperform both general-purpose grasping pipelines and two other online learning algorithms and achieves performance within 5% of the optimal policy within 1000 and 8000 timesteps on average across 46 challenging objects from the Dex-Net adversarial and EGAD! object datasets, respectively. Initial physical experiments suggest that BORGES can improve grasp success rate by 45% over a Dex-Net baseline with just 200 grasp attempts in the real world. See https://tinyurl.com/exp-grasping for supplementary material and videos.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Algorithms ; Decision theory ; Distance learning ; Failure analysis ; Machine learning ; Markov processes ; Policies ; Suction</subject><ispartof>arXiv.org, 2020-11</ispartof><rights>2020. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>780,784</link.rule.ids></links><search><creatorcontrib>Danielczuk, Michael</creatorcontrib><creatorcontrib>Balakrishna, Ashwin</creatorcontrib><creatorcontrib>Brown, Daniel S</creatorcontrib><creatorcontrib>Devgon, Shivin</creatorcontrib><creatorcontrib>Goldberg, Ken</creatorcontrib><title>Exploratory Grasping: Asymptotically Optimal Algorithms for Grasping Challenging Polyhedral Objects</title><title>arXiv.org</title><description>There has been significant recent work on data-driven algorithms for learning general-purpose grasping policies. However, these policies can consistently fail to grasp challenging objects which are significantly out of the distribution of objects in the training data or which have very few high quality grasps. Motivated by such objects, we propose a novel problem setting, Exploratory Grasping, for efficiently discovering reliable grasps on an unknown polyhedral object via sequential grasping, releasing, and toppling. We formalize Exploratory Grasping as a Markov Decision Process, study the theoretical complexity of Exploratory Grasping in the context of reinforcement learning and present an efficient bandit-style algorithm, Bandits for Online Rapid Grasp Exploration Strategy (BORGES), which leverages the structure of the problem to efficiently discover high performing grasps for each object stable pose. BORGES can be used to complement any general-purpose grasping algorithm with any grasp modality (parallel-jaw, suction, multi-fingered, etc) to learn policies for objects in which they exhibit persistent failures. Simulation experiments suggest that BORGES can significantly outperform both general-purpose grasping pipelines and two other online learning algorithms and achieves performance within 5% of the optimal policy within 1000 and 8000 timesteps on average across 46 challenging objects from the Dex-Net adversarial and EGAD! object datasets, respectively. Initial physical experiments suggest that BORGES can improve grasp success rate by 45% over a Dex-Net baseline with just 200 grasp attempts in the real world. See https://tinyurl.com/exp-grasping for supplementary material and videos.</description><subject>Algorithms</subject><subject>Decision theory</subject><subject>Distance learning</subject><subject>Failure analysis</subject><subject>Machine learning</subject><subject>Markov processes</subject><subject>Policies</subject><subject>Suction</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><recordid>eNqNzk8LgjAABfARBEn5HQadhbmp_bmJWN3s0F2WTZ1Mt7YJ7du3IDp3eg_e7_AWIMCExNE-wXgFQmMGhBDOdjhNSQCa8qWE1NRK7eBZU6P41B1hbtyorLS8oUI4WCnLRypgLjqpue1HA1upfx4WvWds6j79KoXr2UN7Xt0H1lizAcuWCsPCb67B9lTeikuktHzOzNh6kLOe_FTjJEPxwb9LyH_qDXBrRwM</recordid><startdate>20201112</startdate><enddate>20201112</enddate><creator>Danielczuk, Michael</creator><creator>Balakrishna, Ashwin</creator><creator>Brown, Daniel S</creator><creator>Devgon, Shivin</creator><creator>Goldberg, Ken</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20201112</creationdate><title>Exploratory Grasping: Asymptotically Optimal Algorithms for Grasping Challenging Polyhedral Objects</title><author>Danielczuk, Michael ; Balakrishna, Ashwin ; Brown, Daniel S ; Devgon, Shivin ; Goldberg, Ken</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_24601926743</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Algorithms</topic><topic>Decision theory</topic><topic>Distance learning</topic><topic>Failure analysis</topic><topic>Machine learning</topic><topic>Markov processes</topic><topic>Policies</topic><topic>Suction</topic><toplevel>online_resources</toplevel><creatorcontrib>Danielczuk, Michael</creatorcontrib><creatorcontrib>Balakrishna, Ashwin</creatorcontrib><creatorcontrib>Brown, Daniel S</creatorcontrib><creatorcontrib>Devgon, Shivin</creatorcontrib><creatorcontrib>Goldberg, Ken</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Access via ProQuest (Open Access)</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Danielczuk, Michael</au><au>Balakrishna, Ashwin</au><au>Brown, Daniel S</au><au>Devgon, Shivin</au><au>Goldberg, Ken</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Exploratory Grasping: Asymptotically Optimal Algorithms for Grasping Challenging Polyhedral Objects</atitle><jtitle>arXiv.org</jtitle><date>2020-11-12</date><risdate>2020</risdate><eissn>2331-8422</eissn><abstract>There has been significant recent work on data-driven algorithms for learning general-purpose grasping policies. However, these policies can consistently fail to grasp challenging objects which are significantly out of the distribution of objects in the training data or which have very few high quality grasps. Motivated by such objects, we propose a novel problem setting, Exploratory Grasping, for efficiently discovering reliable grasps on an unknown polyhedral object via sequential grasping, releasing, and toppling. We formalize Exploratory Grasping as a Markov Decision Process, study the theoretical complexity of Exploratory Grasping in the context of reinforcement learning and present an efficient bandit-style algorithm, Bandits for Online Rapid Grasp Exploration Strategy (BORGES), which leverages the structure of the problem to efficiently discover high performing grasps for each object stable pose. BORGES can be used to complement any general-purpose grasping algorithm with any grasp modality (parallel-jaw, suction, multi-fingered, etc) to learn policies for objects in which they exhibit persistent failures. Simulation experiments suggest that BORGES can significantly outperform both general-purpose grasping pipelines and two other online learning algorithms and achieves performance within 5% of the optimal policy within 1000 and 8000 timesteps on average across 46 challenging objects from the Dex-Net adversarial and EGAD! object datasets, respectively. Initial physical experiments suggest that BORGES can improve grasp success rate by 45% over a Dex-Net baseline with just 200 grasp attempts in the real world. See https://tinyurl.com/exp-grasping for supplementary material and videos.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier EISSN: 2331-8422
ispartof arXiv.org, 2020-11
issn 2331-8422
language eng
recordid cdi_proquest_journals_2460192674
source Free E- Journals
subjects Algorithms
Decision theory
Distance learning
Failure analysis
Machine learning
Markov processes
Policies
Suction
title Exploratory Grasping: Asymptotically Optimal Algorithms for Grasping Challenging Polyhedral Objects
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-26T04%3A11%3A02IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Exploratory%20Grasping:%20Asymptotically%20Optimal%20Algorithms%20for%20Grasping%20Challenging%20Polyhedral%20Objects&rft.jtitle=arXiv.org&rft.au=Danielczuk,%20Michael&rft.date=2020-11-12&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2460192674%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2460192674&rft_id=info:pmid/&rfr_iscdi=true