Imitating and Finetuning Model Predictive Control for Robust and Symmetric Quadrupedal Locomotion

Control of legged robots is a challenging problem that has been investigated by different approaches, such as model-based control and learning algorithms. This work proposes a novel Imitating and Finetuning Model Predictive Control (IFM) framework to take the strengths of both approaches. Our framew...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Youm, Donghoon, Jung, Hyunyoung, Kim, Hyeongjun, Hwangbo, Jemin, Park, Hae-Won, Ha, Sehoon
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 Youm, Donghoon
Jung, Hyunyoung
Kim, Hyeongjun
Hwangbo, Jemin
Park, Hae-Won
Ha, Sehoon
description Control of legged robots is a challenging problem that has been investigated by different approaches, such as model-based control and learning algorithms. This work proposes a novel Imitating and Finetuning Model Predictive Control (IFM) framework to take the strengths of both approaches. Our framework first develops a conventional model predictive controller (MPC) using Differential Dynamic Programming and Raibert heuristic, which serves as an expert policy. Then we train a clone of the MPC using imitation learning to make the controller learnable. Finally, we leverage deep reinforcement learning with limited exploration for further finetuning the policy on more challenging terrains. By conducting comprehensive simulation and hardware experiments, we demonstrate that the proposed IFM framework can significantly improve the performance of the given MPC controller on rough, slippery, and conveyor terrains that require careful coordination of footsteps. We also showcase that IFM can efficiently produce more symmetric, periodic, and energy-efficient gaits compared to Vanilla RL with a minimal burden of reward shaping.
doi_str_mv 10.48550/arxiv.2311.02304
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2311_02304</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2311_02304</sourcerecordid><originalsourceid>FETCH-LOGICAL-a674-f812e8e4d9272e44cba30be625367e95b9005e0b8bc676c057cc4a888bcb48b03</originalsourceid><addsrcrecordid>eNotj8tOwzAUBb1hgQofwAr_QILjR-IsUUShUlB5dB9d2zfIUmJXrlPRv4cWVkezOCMNIXcVK6VWij1A-vbHkouqKhkXTF4T2Mw-Q_bhi0JwdO0D5iWc8TU6nOhbQudt9kekXQw5xYmOMdGPaJZDvlw-T_OMOXlL3xdwadmjg4n20cY5Zh_DDbkaYTrg7f-uyG79tOtein77vOke-wLqRhajrjhqlK7lDUcprQHBDNZcibrBVpmWMYXMaGPrprZMNdZK0PqXjdSGiRW5_9NeGod98jOk03BuHS6t4geP3lBS</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Imitating and Finetuning Model Predictive Control for Robust and Symmetric Quadrupedal Locomotion</title><source>arXiv.org</source><creator>Youm, Donghoon ; Jung, Hyunyoung ; Kim, Hyeongjun ; Hwangbo, Jemin ; Park, Hae-Won ; Ha, Sehoon</creator><creatorcontrib>Youm, Donghoon ; Jung, Hyunyoung ; Kim, Hyeongjun ; Hwangbo, Jemin ; Park, Hae-Won ; Ha, Sehoon</creatorcontrib><description>Control of legged robots is a challenging problem that has been investigated by different approaches, such as model-based control and learning algorithms. This work proposes a novel Imitating and Finetuning Model Predictive Control (IFM) framework to take the strengths of both approaches. Our framework first develops a conventional model predictive controller (MPC) using Differential Dynamic Programming and Raibert heuristic, which serves as an expert policy. Then we train a clone of the MPC using imitation learning to make the controller learnable. Finally, we leverage deep reinforcement learning with limited exploration for further finetuning the policy on more challenging terrains. By conducting comprehensive simulation and hardware experiments, we demonstrate that the proposed IFM framework can significantly improve the performance of the given MPC controller on rough, slippery, and conveyor terrains that require careful coordination of footsteps. We also showcase that IFM can efficiently produce more symmetric, periodic, and energy-efficient gaits compared to Vanilla RL with a minimal burden of reward shaping.</description><identifier>DOI: 10.48550/arxiv.2311.02304</identifier><language>eng</language><subject>Computer Science - Robotics</subject><creationdate>2023-11</creationdate><rights>http://creativecommons.org/licenses/by-nc-sa/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,781,886</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2311.02304$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2311.02304$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Youm, Donghoon</creatorcontrib><creatorcontrib>Jung, Hyunyoung</creatorcontrib><creatorcontrib>Kim, Hyeongjun</creatorcontrib><creatorcontrib>Hwangbo, Jemin</creatorcontrib><creatorcontrib>Park, Hae-Won</creatorcontrib><creatorcontrib>Ha, Sehoon</creatorcontrib><title>Imitating and Finetuning Model Predictive Control for Robust and Symmetric Quadrupedal Locomotion</title><description>Control of legged robots is a challenging problem that has been investigated by different approaches, such as model-based control and learning algorithms. This work proposes a novel Imitating and Finetuning Model Predictive Control (IFM) framework to take the strengths of both approaches. Our framework first develops a conventional model predictive controller (MPC) using Differential Dynamic Programming and Raibert heuristic, which serves as an expert policy. Then we train a clone of the MPC using imitation learning to make the controller learnable. Finally, we leverage deep reinforcement learning with limited exploration for further finetuning the policy on more challenging terrains. By conducting comprehensive simulation and hardware experiments, we demonstrate that the proposed IFM framework can significantly improve the performance of the given MPC controller on rough, slippery, and conveyor terrains that require careful coordination of footsteps. We also showcase that IFM can efficiently produce more symmetric, periodic, and energy-efficient gaits compared to Vanilla RL with a minimal burden of reward shaping.</description><subject>Computer Science - Robotics</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj8tOwzAUBb1hgQofwAr_QILjR-IsUUShUlB5dB9d2zfIUmJXrlPRv4cWVkezOCMNIXcVK6VWij1A-vbHkouqKhkXTF4T2Mw-Q_bhi0JwdO0D5iWc8TU6nOhbQudt9kekXQw5xYmOMdGPaJZDvlw-T_OMOXlL3xdwadmjg4n20cY5Zh_DDbkaYTrg7f-uyG79tOtein77vOke-wLqRhajrjhqlK7lDUcprQHBDNZcibrBVpmWMYXMaGPrprZMNdZK0PqXjdSGiRW5_9NeGod98jOk03BuHS6t4geP3lBS</recordid><startdate>20231103</startdate><enddate>20231103</enddate><creator>Youm, Donghoon</creator><creator>Jung, Hyunyoung</creator><creator>Kim, Hyeongjun</creator><creator>Hwangbo, Jemin</creator><creator>Park, Hae-Won</creator><creator>Ha, Sehoon</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20231103</creationdate><title>Imitating and Finetuning Model Predictive Control for Robust and Symmetric Quadrupedal Locomotion</title><author>Youm, Donghoon ; Jung, Hyunyoung ; Kim, Hyeongjun ; Hwangbo, Jemin ; Park, Hae-Won ; Ha, Sehoon</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a674-f812e8e4d9272e44cba30be625367e95b9005e0b8bc676c057cc4a888bcb48b03</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Computer Science - Robotics</topic><toplevel>online_resources</toplevel><creatorcontrib>Youm, Donghoon</creatorcontrib><creatorcontrib>Jung, Hyunyoung</creatorcontrib><creatorcontrib>Kim, Hyeongjun</creatorcontrib><creatorcontrib>Hwangbo, Jemin</creatorcontrib><creatorcontrib>Park, Hae-Won</creatorcontrib><creatorcontrib>Ha, Sehoon</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Youm, Donghoon</au><au>Jung, Hyunyoung</au><au>Kim, Hyeongjun</au><au>Hwangbo, Jemin</au><au>Park, Hae-Won</au><au>Ha, Sehoon</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Imitating and Finetuning Model Predictive Control for Robust and Symmetric Quadrupedal Locomotion</atitle><date>2023-11-03</date><risdate>2023</risdate><abstract>Control of legged robots is a challenging problem that has been investigated by different approaches, such as model-based control and learning algorithms. This work proposes a novel Imitating and Finetuning Model Predictive Control (IFM) framework to take the strengths of both approaches. Our framework first develops a conventional model predictive controller (MPC) using Differential Dynamic Programming and Raibert heuristic, which serves as an expert policy. Then we train a clone of the MPC using imitation learning to make the controller learnable. Finally, we leverage deep reinforcement learning with limited exploration for further finetuning the policy on more challenging terrains. By conducting comprehensive simulation and hardware experiments, we demonstrate that the proposed IFM framework can significantly improve the performance of the given MPC controller on rough, slippery, and conveyor terrains that require careful coordination of footsteps. We also showcase that IFM can efficiently produce more symmetric, periodic, and energy-efficient gaits compared to Vanilla RL with a minimal burden of reward shaping.</abstract><doi>10.48550/arxiv.2311.02304</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2311.02304
ispartof
issn
language eng
recordid cdi_arxiv_primary_2311_02304
source arXiv.org
subjects Computer Science - Robotics
title Imitating and Finetuning Model Predictive Control for Robust and Symmetric Quadrupedal Locomotion
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-12T00%3A45%3A08IST&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=Imitating%20and%20Finetuning%20Model%20Predictive%20Control%20for%20Robust%20and%20Symmetric%20Quadrupedal%20Locomotion&rft.au=Youm,%20Donghoon&rft.date=2023-11-03&rft_id=info:doi/10.48550/arxiv.2311.02304&rft_dat=%3Carxiv_GOX%3E2311_02304%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