PhyPlan: Generalizable and Rapid Physical Task Planning with Physics Informed Skill Networks for Robot Manipulators

Given the task of positioning a ball-like object to a goal region beyond direct reach, humans can often throw, slide, or rebound objects against the wall to attain the goal. However, enabling robots to reason similarly is non-trivial. Existing methods for physical reasoning are data-hungry and strug...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2024-04
Hauptverfasser: Chopra, Mudit, Barnawal, Abhinav, Vagadia, Harshil, Banerjee, Tamajit, Tuli, Shreshth, Chakraborty, Souvik, Rohan, Paul
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 Chopra, Mudit
Barnawal, Abhinav
Vagadia, Harshil
Banerjee, Tamajit
Tuli, Shreshth
Chakraborty, Souvik
Rohan, Paul
description Given the task of positioning a ball-like object to a goal region beyond direct reach, humans can often throw, slide, or rebound objects against the wall to attain the goal. However, enabling robots to reason similarly is non-trivial. Existing methods for physical reasoning are data-hungry and struggle with complexity and uncertainty inherent in the real world. This paper presents PhyPlan, a novel physics-informed planning framework that combines physics-informed neural networks (PINNs) with modified Monte Carlo Tree Search (MCTS) to enable embodied agents to perform dynamic physical tasks. PhyPlan leverages PINNs to simulate and predict outcomes of actions in a fast and accurate manner and uses MCTS for planning. It dynamically determines whether to consult a PINN-based simulator (coarse but fast) or engage directly with the actual environment (fine but slow) to determine optimal policy. Given an unseen task, PhyPlan can infer the sequence of actions and learn the latent parameters, resulting in a generalizable approach that can rapidly learn to perform novel physical tasks. Evaluation with robots in simulated 3D environments demonstrates the ability of our approach to solve 3D-physical reasoning tasks involving the composition of dynamic skills. Quantitatively, PhyPlan excels in several aspects: (i) it achieves lower regret when learning novel tasks compared to the state-of-the-art, (ii) it expedites skill learning and enhances the speed of physical reasoning, (iii) it demonstrates higher data efficiency compared to a physics un-informed approach.
format Article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_3064392663</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3064392663</sourcerecordid><originalsourceid>FETCH-proquest_journals_30643926633</originalsourceid><addsrcrecordid>eNqNyksKwjAUheEgCIq6hwuOCzWx9TEVXwOlVOdya6PGxqTmphRdvRVcgKMD5_9arMuFGAXTMecdNiC6h2HI4wmPItFllNxeiUYzh7U00qFWb8y0BDQ5pFiqHBpA6owajkgFfK1R5gq18rdfI9iai3UPmcOhUFrDXvrauoKgeSG1mfWwQ6PKSqO3jvqsfUFNcvDbHhuulsfFJiidfVaS_OluK2eadBJhPBYzHsdC_Kc-pBNL3A</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3064392663</pqid></control><display><type>article</type><title>PhyPlan: Generalizable and Rapid Physical Task Planning with Physics Informed Skill Networks for Robot Manipulators</title><source>Free E- Journals</source><creator>Chopra, Mudit ; Barnawal, Abhinav ; Vagadia, Harshil ; Banerjee, Tamajit ; Tuli, Shreshth ; Chakraborty, Souvik ; Rohan, Paul</creator><creatorcontrib>Chopra, Mudit ; Barnawal, Abhinav ; Vagadia, Harshil ; Banerjee, Tamajit ; Tuli, Shreshth ; Chakraborty, Souvik ; Rohan, Paul</creatorcontrib><description>Given the task of positioning a ball-like object to a goal region beyond direct reach, humans can often throw, slide, or rebound objects against the wall to attain the goal. However, enabling robots to reason similarly is non-trivial. Existing methods for physical reasoning are data-hungry and struggle with complexity and uncertainty inherent in the real world. This paper presents PhyPlan, a novel physics-informed planning framework that combines physics-informed neural networks (PINNs) with modified Monte Carlo Tree Search (MCTS) to enable embodied agents to perform dynamic physical tasks. PhyPlan leverages PINNs to simulate and predict outcomes of actions in a fast and accurate manner and uses MCTS for planning. It dynamically determines whether to consult a PINN-based simulator (coarse but fast) or engage directly with the actual environment (fine but slow) to determine optimal policy. Given an unseen task, PhyPlan can infer the sequence of actions and learn the latent parameters, resulting in a generalizable approach that can rapidly learn to perform novel physical tasks. Evaluation with robots in simulated 3D environments demonstrates the ability of our approach to solve 3D-physical reasoning tasks involving the composition of dynamic skills. Quantitatively, PhyPlan excels in several aspects: (i) it achieves lower regret when learning novel tasks compared to the state-of-the-art, (ii) it expedites skill learning and enhances the speed of physical reasoning, (iii) it demonstrates higher data efficiency compared to a physics un-informed approach.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Decision theory ; Learning ; Neural networks ; Physics ; Reasoning ; Robot arms ; Task planning (robotics)</subject><ispartof>arXiv.org, 2024-04</ispartof><rights>2024. This work is published under http://creativecommons.org/licenses/by-nc-nd/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>776,780</link.rule.ids></links><search><creatorcontrib>Chopra, Mudit</creatorcontrib><creatorcontrib>Barnawal, Abhinav</creatorcontrib><creatorcontrib>Vagadia, Harshil</creatorcontrib><creatorcontrib>Banerjee, Tamajit</creatorcontrib><creatorcontrib>Tuli, Shreshth</creatorcontrib><creatorcontrib>Chakraborty, Souvik</creatorcontrib><creatorcontrib>Rohan, Paul</creatorcontrib><title>PhyPlan: Generalizable and Rapid Physical Task Planning with Physics Informed Skill Networks for Robot Manipulators</title><title>arXiv.org</title><description>Given the task of positioning a ball-like object to a goal region beyond direct reach, humans can often throw, slide, or rebound objects against the wall to attain the goal. However, enabling robots to reason similarly is non-trivial. Existing methods for physical reasoning are data-hungry and struggle with complexity and uncertainty inherent in the real world. This paper presents PhyPlan, a novel physics-informed planning framework that combines physics-informed neural networks (PINNs) with modified Monte Carlo Tree Search (MCTS) to enable embodied agents to perform dynamic physical tasks. PhyPlan leverages PINNs to simulate and predict outcomes of actions in a fast and accurate manner and uses MCTS for planning. It dynamically determines whether to consult a PINN-based simulator (coarse but fast) or engage directly with the actual environment (fine but slow) to determine optimal policy. Given an unseen task, PhyPlan can infer the sequence of actions and learn the latent parameters, resulting in a generalizable approach that can rapidly learn to perform novel physical tasks. Evaluation with robots in simulated 3D environments demonstrates the ability of our approach to solve 3D-physical reasoning tasks involving the composition of dynamic skills. Quantitatively, PhyPlan excels in several aspects: (i) it achieves lower regret when learning novel tasks compared to the state-of-the-art, (ii) it expedites skill learning and enhances the speed of physical reasoning, (iii) it demonstrates higher data efficiency compared to a physics un-informed approach.</description><subject>Decision theory</subject><subject>Learning</subject><subject>Neural networks</subject><subject>Physics</subject><subject>Reasoning</subject><subject>Robot arms</subject><subject>Task planning (robotics)</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>BENPR</sourceid><recordid>eNqNyksKwjAUheEgCIq6hwuOCzWx9TEVXwOlVOdya6PGxqTmphRdvRVcgKMD5_9arMuFGAXTMecdNiC6h2HI4wmPItFllNxeiUYzh7U00qFWb8y0BDQ5pFiqHBpA6owajkgFfK1R5gq18rdfI9iai3UPmcOhUFrDXvrauoKgeSG1mfWwQ6PKSqO3jvqsfUFNcvDbHhuulsfFJiidfVaS_OluK2eadBJhPBYzHsdC_Kc-pBNL3A</recordid><startdate>20240422</startdate><enddate>20240422</enddate><creator>Chopra, Mudit</creator><creator>Barnawal, Abhinav</creator><creator>Vagadia, Harshil</creator><creator>Banerjee, Tamajit</creator><creator>Tuli, Shreshth</creator><creator>Chakraborty, Souvik</creator><creator>Rohan, Paul</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>20240422</creationdate><title>PhyPlan: Generalizable and Rapid Physical Task Planning with Physics Informed Skill Networks for Robot Manipulators</title><author>Chopra, Mudit ; Barnawal, Abhinav ; Vagadia, Harshil ; Banerjee, Tamajit ; Tuli, Shreshth ; Chakraborty, Souvik ; Rohan, Paul</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_30643926633</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Decision theory</topic><topic>Learning</topic><topic>Neural networks</topic><topic>Physics</topic><topic>Reasoning</topic><topic>Robot arms</topic><topic>Task planning (robotics)</topic><toplevel>online_resources</toplevel><creatorcontrib>Chopra, Mudit</creatorcontrib><creatorcontrib>Barnawal, Abhinav</creatorcontrib><creatorcontrib>Vagadia, Harshil</creatorcontrib><creatorcontrib>Banerjee, Tamajit</creatorcontrib><creatorcontrib>Tuli, Shreshth</creatorcontrib><creatorcontrib>Chakraborty, Souvik</creatorcontrib><creatorcontrib>Rohan, Paul</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; Engineering Collection</collection><collection>ProQuest Central (Alumni)</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</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>ProQuest Engineering Database</collection><collection>Publicly Available Content Database</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>Chopra, Mudit</au><au>Barnawal, Abhinav</au><au>Vagadia, Harshil</au><au>Banerjee, Tamajit</au><au>Tuli, Shreshth</au><au>Chakraborty, Souvik</au><au>Rohan, Paul</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>PhyPlan: Generalizable and Rapid Physical Task Planning with Physics Informed Skill Networks for Robot Manipulators</atitle><jtitle>arXiv.org</jtitle><date>2024-04-22</date><risdate>2024</risdate><eissn>2331-8422</eissn><abstract>Given the task of positioning a ball-like object to a goal region beyond direct reach, humans can often throw, slide, or rebound objects against the wall to attain the goal. However, enabling robots to reason similarly is non-trivial. Existing methods for physical reasoning are data-hungry and struggle with complexity and uncertainty inherent in the real world. This paper presents PhyPlan, a novel physics-informed planning framework that combines physics-informed neural networks (PINNs) with modified Monte Carlo Tree Search (MCTS) to enable embodied agents to perform dynamic physical tasks. PhyPlan leverages PINNs to simulate and predict outcomes of actions in a fast and accurate manner and uses MCTS for planning. It dynamically determines whether to consult a PINN-based simulator (coarse but fast) or engage directly with the actual environment (fine but slow) to determine optimal policy. Given an unseen task, PhyPlan can infer the sequence of actions and learn the latent parameters, resulting in a generalizable approach that can rapidly learn to perform novel physical tasks. Evaluation with robots in simulated 3D environments demonstrates the ability of our approach to solve 3D-physical reasoning tasks involving the composition of dynamic skills. Quantitatively, PhyPlan excels in several aspects: (i) it achieves lower regret when learning novel tasks compared to the state-of-the-art, (ii) it expedites skill learning and enhances the speed of physical reasoning, (iii) it demonstrates higher data efficiency compared to a physics un-informed approach.</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, 2024-04
issn 2331-8422
language eng
recordid cdi_proquest_journals_3064392663
source Free E- Journals
subjects Decision theory
Learning
Neural networks
Physics
Reasoning
Robot arms
Task planning (robotics)
title PhyPlan: Generalizable and Rapid Physical Task Planning with Physics Informed Skill Networks for Robot Manipulators
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-07T16%3A56%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=PhyPlan:%20Generalizable%20and%20Rapid%20Physical%20Task%20Planning%20with%20Physics%20Informed%20Skill%20Networks%20for%20Robot%20Manipulators&rft.jtitle=arXiv.org&rft.au=Chopra,%20Mudit&rft.date=2024-04-22&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E3064392663%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3064392663&rft_id=info:pmid/&rfr_iscdi=true