Real-time Model Predictive Control and System Identification Using Differentiable Physics Simulation

Developing robot controllers in a simulated environment is advantageous but transferring the controllers to the target environment presents challenges, often referred to as the "sim-to-real gap". We present a method for continuous improvement of modeling and control after deploying the rob...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chen, Sirui, Werling, Keenon, Wu, Albert, Liu, C. Karen
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Chen, Sirui
Werling, Keenon
Wu, Albert
Liu, C. Karen
description Developing robot controllers in a simulated environment is advantageous but transferring the controllers to the target environment presents challenges, often referred to as the "sim-to-real gap". We present a method for continuous improvement of modeling and control after deploying the robot to a dynamically-changing target environment. We develop a differentiable physics simulation framework that performs online system identification and optimal control simultaneously, using the incoming observations from the target environment in real time. To ensure robust system identification against noisy observations, we devise an algorithm to assess the confidence of our estimated parameters, using numerical analysis of the dynamic equations. To ensure real-time optimal control, we adaptively schedule the optimization window in the future so that the optimized actions can be replenished faster than they are consumed, while staying as up-to-date with new sensor information as possible. The constant re-planning based on a constantly improved model allows the robot to swiftly adapt to the changing environment and utilize real-world data in the most sample-efficient way. Thanks to a fast differentiable physics simulator, the optimization for both system identification and control can be solved efficiently for robots operating in real time. We demonstrate our method on a set of examples in simulation and show that our results are favorable compared to baseline methods.
doi_str_mv 10.48550/arxiv.2202.09834
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2202_09834</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2202_09834</sourcerecordid><originalsourceid>FETCH-LOGICAL-a674-89b1b117c0badf8350644c6b9fc69af70147cfd8762c3eb36e4589a8023b7ed93</originalsourceid><addsrcrecordid>eNotz8tOwzAUBNBsWKDCB7DCP5DgxI4fSxRelVpR0bKO_LiGKzkJckxF_x41sJrFjEY6RXFT04qrtqV3Jv3gsWoa2lRUK8YvC_8GJpYZByDbyUMkuwQeXcYjkG4ac5oiMaMn-9OcYSBrD2PGgM5knEbyPuP4QR4wBEjnwtgIZPd5mtHNZI_Dd1x2V8VFMHGG6_9cFYenx0P3Um5en9fd_aY0QvJSaVvbupaOWuODYi0VnDthdXBCmyBpzaULXknROAaWCeCt0kbRhlkJXrNVcft3uzD7r4SDSaf-zO0XLvsFotpRfw</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Real-time Model Predictive Control and System Identification Using Differentiable Physics Simulation</title><source>arXiv.org</source><creator>Chen, Sirui ; Werling, Keenon ; Wu, Albert ; Liu, C. Karen</creator><creatorcontrib>Chen, Sirui ; Werling, Keenon ; Wu, Albert ; Liu, C. Karen</creatorcontrib><description>Developing robot controllers in a simulated environment is advantageous but transferring the controllers to the target environment presents challenges, often referred to as the "sim-to-real gap". We present a method for continuous improvement of modeling and control after deploying the robot to a dynamically-changing target environment. We develop a differentiable physics simulation framework that performs online system identification and optimal control simultaneously, using the incoming observations from the target environment in real time. To ensure robust system identification against noisy observations, we devise an algorithm to assess the confidence of our estimated parameters, using numerical analysis of the dynamic equations. To ensure real-time optimal control, we adaptively schedule the optimization window in the future so that the optimized actions can be replenished faster than they are consumed, while staying as up-to-date with new sensor information as possible. The constant re-planning based on a constantly improved model allows the robot to swiftly adapt to the changing environment and utilize real-world data in the most sample-efficient way. Thanks to a fast differentiable physics simulator, the optimization for both system identification and control can be solved efficiently for robots operating in real time. We demonstrate our method on a set of examples in simulation and show that our results are favorable compared to baseline methods.</description><identifier>DOI: 10.48550/arxiv.2202.09834</identifier><language>eng</language><subject>Computer Science - Graphics ; Computer Science - Robotics</subject><creationdate>2022-02</creationdate><rights>http://creativecommons.org/licenses/by-nc-nd/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,777,882</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2202.09834$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2202.09834$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Chen, Sirui</creatorcontrib><creatorcontrib>Werling, Keenon</creatorcontrib><creatorcontrib>Wu, Albert</creatorcontrib><creatorcontrib>Liu, C. Karen</creatorcontrib><title>Real-time Model Predictive Control and System Identification Using Differentiable Physics Simulation</title><description>Developing robot controllers in a simulated environment is advantageous but transferring the controllers to the target environment presents challenges, often referred to as the "sim-to-real gap". We present a method for continuous improvement of modeling and control after deploying the robot to a dynamically-changing target environment. We develop a differentiable physics simulation framework that performs online system identification and optimal control simultaneously, using the incoming observations from the target environment in real time. To ensure robust system identification against noisy observations, we devise an algorithm to assess the confidence of our estimated parameters, using numerical analysis of the dynamic equations. To ensure real-time optimal control, we adaptively schedule the optimization window in the future so that the optimized actions can be replenished faster than they are consumed, while staying as up-to-date with new sensor information as possible. The constant re-planning based on a constantly improved model allows the robot to swiftly adapt to the changing environment and utilize real-world data in the most sample-efficient way. Thanks to a fast differentiable physics simulator, the optimization for both system identification and control can be solved efficiently for robots operating in real time. We demonstrate our method on a set of examples in simulation and show that our results are favorable compared to baseline methods.</description><subject>Computer Science - Graphics</subject><subject>Computer Science - Robotics</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotz8tOwzAUBNBsWKDCB7DCP5DgxI4fSxRelVpR0bKO_LiGKzkJckxF_x41sJrFjEY6RXFT04qrtqV3Jv3gsWoa2lRUK8YvC_8GJpYZByDbyUMkuwQeXcYjkG4ac5oiMaMn-9OcYSBrD2PGgM5knEbyPuP4QR4wBEjnwtgIZPd5mtHNZI_Dd1x2V8VFMHGG6_9cFYenx0P3Um5en9fd_aY0QvJSaVvbupaOWuODYi0VnDthdXBCmyBpzaULXknROAaWCeCt0kbRhlkJXrNVcft3uzD7r4SDSaf-zO0XLvsFotpRfw</recordid><startdate>20220220</startdate><enddate>20220220</enddate><creator>Chen, Sirui</creator><creator>Werling, Keenon</creator><creator>Wu, Albert</creator><creator>Liu, C. Karen</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20220220</creationdate><title>Real-time Model Predictive Control and System Identification Using Differentiable Physics Simulation</title><author>Chen, Sirui ; Werling, Keenon ; Wu, Albert ; Liu, C. Karen</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a674-89b1b117c0badf8350644c6b9fc69af70147cfd8762c3eb36e4589a8023b7ed93</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Computer Science - Graphics</topic><topic>Computer Science - Robotics</topic><toplevel>online_resources</toplevel><creatorcontrib>Chen, Sirui</creatorcontrib><creatorcontrib>Werling, Keenon</creatorcontrib><creatorcontrib>Wu, Albert</creatorcontrib><creatorcontrib>Liu, C. Karen</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Chen, Sirui</au><au>Werling, Keenon</au><au>Wu, Albert</au><au>Liu, C. Karen</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Real-time Model Predictive Control and System Identification Using Differentiable Physics Simulation</atitle><date>2022-02-20</date><risdate>2022</risdate><abstract>Developing robot controllers in a simulated environment is advantageous but transferring the controllers to the target environment presents challenges, often referred to as the "sim-to-real gap". We present a method for continuous improvement of modeling and control after deploying the robot to a dynamically-changing target environment. We develop a differentiable physics simulation framework that performs online system identification and optimal control simultaneously, using the incoming observations from the target environment in real time. To ensure robust system identification against noisy observations, we devise an algorithm to assess the confidence of our estimated parameters, using numerical analysis of the dynamic equations. To ensure real-time optimal control, we adaptively schedule the optimization window in the future so that the optimized actions can be replenished faster than they are consumed, while staying as up-to-date with new sensor information as possible. The constant re-planning based on a constantly improved model allows the robot to swiftly adapt to the changing environment and utilize real-world data in the most sample-efficient way. Thanks to a fast differentiable physics simulator, the optimization for both system identification and control can be solved efficiently for robots operating in real time. We demonstrate our method on a set of examples in simulation and show that our results are favorable compared to baseline methods.</abstract><doi>10.48550/arxiv.2202.09834</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2202.09834
ispartof
issn
language eng
recordid cdi_arxiv_primary_2202_09834
source arXiv.org
subjects Computer Science - Graphics
Computer Science - Robotics
title Real-time Model Predictive Control and System Identification Using Differentiable Physics Simulation
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-19T08%3A53%3A44IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Real-time%20Model%20Predictive%20Control%20and%20System%20Identification%20Using%20Differentiable%20Physics%20Simulation&rft.au=Chen,%20Sirui&rft.date=2022-02-20&rft_id=info:doi/10.48550/arxiv.2202.09834&rft_dat=%3Carxiv_GOX%3E2202_09834%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true