No Cover Image

E-Thesis 295 views 49 downloads

Real-Time Hybrid Visual Servoing of a Redundant Manipulator via Deep Reinforcement Learning / Alex Williams

Swansea University Author: Alex Williams

  • Williams_Alexander_J_MSc_Research_Thesis_Final_Redacted_Signature.pdf

    PDF | E-Thesis – open access

    Copyright: The author, Alexander J. Williams, 2023. Released under the terms of a Creative Commons Attribution-Only (CC-BY) License. Third party content is excluded for use under the license terms.

    Download (7.17MB)

Abstract

Fixtureless assembly may be necessary in some manufacturing tasks and environ-ments due to various constraints but poses challenges for automation due to non-deterministic characteristics not favoured by traditional approaches to industrial au-tomation. Visual servoing methods of robotic control cou...

Full description

Published: Swansea 2023
Institution: Swansea University
Degree level: Master of Research
Degree name: MSc by Research
Supervisor: Griffiths, Christian ; Cameron, Ian
URI: https://cronfa.swan.ac.uk/Record/cronfa62598
Tags: Add Tag
artificial intelligence 1
collaborative robots 1
reinforcement learning 1
robotics 1
first_indexed 2023-02-07T11:43:27Z
last_indexed 2023-02-08T04:17:09Z
id cronfa62598
recordtype RisThesis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2023-02-07T11:57:35.7522672</datestamp><bib-version>v2</bib-version><id>62598</id><entry>2023-02-07</entry><title>Real-Time Hybrid Visual Servoing of a Redundant Manipulator via Deep Reinforcement Learning</title><swanseaauthors><author><sid>7fead5851d72ae17b6936afd3ee4533c</sid><ORCID>0000-0003-2387-6876</ORCID><firstname>Alex</firstname><surname>Williams</surname><name>Alex Williams</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2023-02-07</date><deptcode>AERO</deptcode><abstract>Fixtureless assembly may be necessary in some manufacturing tasks and environ-ments due to various constraints but poses challenges for automation due to non-deterministic characteristics not favoured by traditional approaches to industrial au-tomation. Visual servoing methods of robotic control could be effective for sensitive manipulation tasks where the desired end-effector pose can be ascertained via visual cues. Visual data is complex and computationally expensive to process but deep reinforcement learning has shown promise for robotic control in vision-based manipu-lation tasks. However, these methods are rarely used in industry due to the resources and expertise required to develop application-specific systems and prohibitive train-ing costs. Training reinforcement learning models in simulated environments offers a number of benefits for the development of robust robotic control algorithms by reducing training time and costs, and providing repeatable benchmarks for which algorithms can be tested, developed and eventually deployed on real robotic control environments. In this work, we present a new simulated reinforcement learning envi-ronment for developing accurate robotic manipulation control systems in fixtureless environments. Our environment incorporates a contemporary collaborative industrial robot, the KUKA LBR iiwa, with the goal of positioning its end effector in a generic fixtureless environment based on a visual cue. Observational inputs are comprised of the robotic joint positions and velocities, as well as two cameras, whose positioning reflect hybrid visual servoing with one camera attached to the robotic end-effector, and another observing the workspace respectively. We propose a state-of-the-art deep reinforcement learning approach to solving the task environment and make prelimi-nary assessments of the efficacy of this approach to hybrid visual servoing methods for the defined problem environment. We also conduct a series of experiments ex-ploring the hyperparameter space in the proposed reinforcement learning method. Although we could not prove the efficacy of a deep reinforcement approach to solving the task environment with our initial results, we remain confident that such an ap-proach could be feasible to solving this industrial manufacturing challenge and that our contributions in this work in terms of the novel software provide a good basis for the exploration of reinforcement learning approaches to hybrid visual servoing in accurate manufacturing contexts.</abstract><type>E-Thesis</type><journal/><volume/><journalNumber/><paginationStart/><paginationEnd/><publisher/><placeOfPublication>Swansea</placeOfPublication><isbnPrint/><isbnElectronic/><issnPrint/><issnElectronic/><keywords>Artificial Intelligence, Robotics, Reinforcement Learning, Machine Learning, Manufacturing Systems Engineering, Aerospace Manufacturing</keywords><publishedDay>1</publishedDay><publishedMonth>2</publishedMonth><publishedYear>2023</publishedYear><publishedDate>2023-02-01</publishedDate><doi/><url/><notes>Research data URL:https://github.com/alexjameswilliams/KukaHybridVisualServoinghttps://www.comet.ml/alexjameswilliams/ kukahybridvs/ ; ORCiD identifier: https://orcid.org/0000-0003-2387-6876</notes><college>COLLEGE NANME</college><department>Aerospace Engineering</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>AERO</DepartmentCode><institution>Swansea University</institution><supervisor>Griffiths, Christian ; Cameron, Ian</supervisor><degreelevel>Master of Research</degreelevel><degreename>MSc by Research</degreename><degreesponsorsfunders>ASTUTE 2020</degreesponsorsfunders><apcterm/><funders/><projectreference/><lastEdited>2023-02-07T11:57:35.7522672</lastEdited><Created>2023-02-07T11:13:31.1539224</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Aerospace Engineering</level></path><authors><author><firstname>Alex</firstname><surname>Williams</surname><orcid>0000-0003-2387-6876</orcid><order>1</order></author></authors><documents><document><filename>62598__26504__2afff3d9de1c467291f5c667c59fbc0e.pdf</filename><originalFilename>Williams_Alexander_J_MSc_Research_Thesis_Final_Redacted_Signature.pdf</originalFilename><uploaded>2023-02-07T11:49:00.6564892</uploaded><type>Output</type><contentLength>7523266</contentLength><contentType>application/pdf</contentType><version>E-Thesis &#x2013; open access</version><cronfaStatus>true</cronfaStatus><documentNotes>Copyright: The author, Alexander J. Williams, 2023. Released under the terms of a Creative Commons Attribution-Only (CC-BY) License. Third party content is excluded for use under the license terms.</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>https://creativecommons.org/licenses/by/4.0/</licence></document></documents><OutputDurs/></rfc1807>
spelling 2023-02-07T11:57:35.7522672 v2 62598 2023-02-07 Real-Time Hybrid Visual Servoing of a Redundant Manipulator via Deep Reinforcement Learning 7fead5851d72ae17b6936afd3ee4533c 0000-0003-2387-6876 Alex Williams Alex Williams true false 2023-02-07 AERO Fixtureless assembly may be necessary in some manufacturing tasks and environ-ments due to various constraints but poses challenges for automation due to non-deterministic characteristics not favoured by traditional approaches to industrial au-tomation. Visual servoing methods of robotic control could be effective for sensitive manipulation tasks where the desired end-effector pose can be ascertained via visual cues. Visual data is complex and computationally expensive to process but deep reinforcement learning has shown promise for robotic control in vision-based manipu-lation tasks. However, these methods are rarely used in industry due to the resources and expertise required to develop application-specific systems and prohibitive train-ing costs. Training reinforcement learning models in simulated environments offers a number of benefits for the development of robust robotic control algorithms by reducing training time and costs, and providing repeatable benchmarks for which algorithms can be tested, developed and eventually deployed on real robotic control environments. In this work, we present a new simulated reinforcement learning envi-ronment for developing accurate robotic manipulation control systems in fixtureless environments. Our environment incorporates a contemporary collaborative industrial robot, the KUKA LBR iiwa, with the goal of positioning its end effector in a generic fixtureless environment based on a visual cue. Observational inputs are comprised of the robotic joint positions and velocities, as well as two cameras, whose positioning reflect hybrid visual servoing with one camera attached to the robotic end-effector, and another observing the workspace respectively. We propose a state-of-the-art deep reinforcement learning approach to solving the task environment and make prelimi-nary assessments of the efficacy of this approach to hybrid visual servoing methods for the defined problem environment. We also conduct a series of experiments ex-ploring the hyperparameter space in the proposed reinforcement learning method. Although we could not prove the efficacy of a deep reinforcement approach to solving the task environment with our initial results, we remain confident that such an ap-proach could be feasible to solving this industrial manufacturing challenge and that our contributions in this work in terms of the novel software provide a good basis for the exploration of reinforcement learning approaches to hybrid visual servoing in accurate manufacturing contexts. E-Thesis Swansea Artificial Intelligence, Robotics, Reinforcement Learning, Machine Learning, Manufacturing Systems Engineering, Aerospace Manufacturing 1 2 2023 2023-02-01 Research data URL:https://github.com/alexjameswilliams/KukaHybridVisualServoinghttps://www.comet.ml/alexjameswilliams/ kukahybridvs/ ; ORCiD identifier: https://orcid.org/0000-0003-2387-6876 COLLEGE NANME Aerospace Engineering COLLEGE CODE AERO Swansea University Griffiths, Christian ; Cameron, Ian Master of Research MSc by Research ASTUTE 2020 2023-02-07T11:57:35.7522672 2023-02-07T11:13:31.1539224 Faculty of Science and Engineering School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Aerospace Engineering Alex Williams 0000-0003-2387-6876 1 62598__26504__2afff3d9de1c467291f5c667c59fbc0e.pdf Williams_Alexander_J_MSc_Research_Thesis_Final_Redacted_Signature.pdf 2023-02-07T11:49:00.6564892 Output 7523266 application/pdf E-Thesis – open access true Copyright: The author, Alexander J. Williams, 2023. Released under the terms of a Creative Commons Attribution-Only (CC-BY) License. Third party content is excluded for use under the license terms. true eng https://creativecommons.org/licenses/by/4.0/
title Real-Time Hybrid Visual Servoing of a Redundant Manipulator via Deep Reinforcement Learning
spellingShingle Real-Time Hybrid Visual Servoing of a Redundant Manipulator via Deep Reinforcement Learning
Alex Williams
title_short Real-Time Hybrid Visual Servoing of a Redundant Manipulator via Deep Reinforcement Learning
title_full Real-Time Hybrid Visual Servoing of a Redundant Manipulator via Deep Reinforcement Learning
title_fullStr Real-Time Hybrid Visual Servoing of a Redundant Manipulator via Deep Reinforcement Learning
title_full_unstemmed Real-Time Hybrid Visual Servoing of a Redundant Manipulator via Deep Reinforcement Learning
title_sort Real-Time Hybrid Visual Servoing of a Redundant Manipulator via Deep Reinforcement Learning
author_id_str_mv 7fead5851d72ae17b6936afd3ee4533c
author_id_fullname_str_mv 7fead5851d72ae17b6936afd3ee4533c_***_Alex Williams
author Alex Williams
author2 Alex Williams
format E-Thesis
publishDate 2023
institution Swansea University
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Aerospace Engineering{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Aerospace Engineering
document_store_str 1
active_str 0
description Fixtureless assembly may be necessary in some manufacturing tasks and environ-ments due to various constraints but poses challenges for automation due to non-deterministic characteristics not favoured by traditional approaches to industrial au-tomation. Visual servoing methods of robotic control could be effective for sensitive manipulation tasks where the desired end-effector pose can be ascertained via visual cues. Visual data is complex and computationally expensive to process but deep reinforcement learning has shown promise for robotic control in vision-based manipu-lation tasks. However, these methods are rarely used in industry due to the resources and expertise required to develop application-specific systems and prohibitive train-ing costs. Training reinforcement learning models in simulated environments offers a number of benefits for the development of robust robotic control algorithms by reducing training time and costs, and providing repeatable benchmarks for which algorithms can be tested, developed and eventually deployed on real robotic control environments. In this work, we present a new simulated reinforcement learning envi-ronment for developing accurate robotic manipulation control systems in fixtureless environments. Our environment incorporates a contemporary collaborative industrial robot, the KUKA LBR iiwa, with the goal of positioning its end effector in a generic fixtureless environment based on a visual cue. Observational inputs are comprised of the robotic joint positions and velocities, as well as two cameras, whose positioning reflect hybrid visual servoing with one camera attached to the robotic end-effector, and another observing the workspace respectively. We propose a state-of-the-art deep reinforcement learning approach to solving the task environment and make prelimi-nary assessments of the efficacy of this approach to hybrid visual servoing methods for the defined problem environment. We also conduct a series of experiments ex-ploring the hyperparameter space in the proposed reinforcement learning method. Although we could not prove the efficacy of a deep reinforcement approach to solving the task environment with our initial results, we remain confident that such an ap-proach could be feasible to solving this industrial manufacturing challenge and that our contributions in this work in terms of the novel software provide a good basis for the exploration of reinforcement learning approaches to hybrid visual servoing in accurate manufacturing contexts.
published_date 2023-02-01T04:22:20Z
_version_ 1763754472062844928
score 11.016302

Similar Items