서브메뉴
검색
Control and Motion Planning for a Low-Inertia Multi-DoF Robotic Manipulator With Proprioceptive Actuators for Dynamic Manipulation
Control and Motion Planning for a Low-Inertia Multi-DoF Robotic Manipulator With Proprioceptive Actuators for Dynamic Manipulation
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211153126
- ISBN
- 9798346857839
- DDC
- 629.8
- 저자명
- Noh, Donghun.
- 서명/저자
- Control and Motion Planning for a Low-Inertia Multi-DoF Robotic Manipulator With Proprioceptive Actuators for Dynamic Manipulation
- 발행사항
- [Sl] : University of California, Los Angeles, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 148 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-06, Section: B.
- 주기사항
- Advisor: Hong, Dennis.
- 학위논문주기
- Thesis (Ph.D.)--University of California, Los Angeles, 2024.
- 초록/해제
- 요약Robotic manipulators are increasingly required to operate in dynamic, human-centric environments, undertaking tasks that range from delicate grasping to high-speed, high-impact manipulation. While conventional high-gear-ratio manipulators and position control strategies perform admirably in structured industrial contexts, they often prove inadequate for these more complex, unstructured domains. This dissertation addresses these shortcomings by integrating novel hardware, control, motion planning, and machine learning methodologies to realize more adaptable and robust robotic manipulation.Central to this work is the deployment of a proprioceptive manipulator, whose inherent backdrivability, high speed, and precise torque control distinguish them from conventional manipulators. Their low-inertia, lightweight design facilitates safer and more responsive robot-environment interactions. To fully exploit these advantages, we develop control methodologies that emphasize torque control rather than position control. In parallel, we introduce motion planning strategies that minimize jerk and incorporate dynamic constraints, thereby producing trajectories that are both efficient and smooth.This dissertation further enhances adaptability through machine-learning approaches tailored for friction-aware grasping and dynamic manipulation. By constructing a friction coefficient dataset and training regression models, the system can efficiently adjust grasping forces in real-time. Additionally, ongoing research in imitation learning, drawing on human demonstrations and incorporating both visual data and physical models, shows promise in enabling the robot to eventually replicate human-level skill in complex, dynamically varying tasks.Comprehensive evaluations include successful demonstrations of autonomous cooking and friction-aware grasping, as well as ongoing research into dynamic box-receiving tasks. While further validation and optimization remain active areas of study, results so far suggest that, when integrated with torque control, optimized motion planning, and model-based data-driven learning, the proposed proprioceptive manipulator can significantly enhance manipulation speed, stability, energy efficiency, and compliance. These findings provide a strong foundation for more intuitive, versatile, and collaborative robotic systems, better suited to the demands of real-world, human-centered environments.
- 일반주제명
- Robotics
- 일반주제명
- Computer engineering
- 일반주제명
- Mechanical engineering
- 키워드
- Machine learning
- 키워드
- Optimal control
- 기타저자
- University of California, Los Angeles Mechanical Engineering 0330
- 기본자료저록
- Dissertations Abstracts International. 86-06B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017165120
■00520250211153126
■006m o d
■007cr#unu||||||||
■020 ▼a9798346857839
■035 ▼a(MiAaPQ)AAI31765141
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a629.8
■1001 ▼aNoh, Donghun.
■24510▼aControl and Motion Planning for a Low-Inertia Multi-DoF Robotic Manipulator With Proprioceptive Actuators for Dynamic Manipulation
■260 ▼a[Sl]▼bUniversity of California, Los Angeles▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a148 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-06, Section: B.
■500 ▼aAdvisor: Hong, Dennis.
■5021 ▼aThesis (Ph.D.)--University of California, Los Angeles, 2024.
■520 ▼aRobotic manipulators are increasingly required to operate in dynamic, human-centric environments, undertaking tasks that range from delicate grasping to high-speed, high-impact manipulation. While conventional high-gear-ratio manipulators and position control strategies perform admirably in structured industrial contexts, they often prove inadequate for these more complex, unstructured domains. This dissertation addresses these shortcomings by integrating novel hardware, control, motion planning, and machine learning methodologies to realize more adaptable and robust robotic manipulation.Central to this work is the deployment of a proprioceptive manipulator, whose inherent backdrivability, high speed, and precise torque control distinguish them from conventional manipulators. Their low-inertia, lightweight design facilitates safer and more responsive robot-environment interactions. To fully exploit these advantages, we develop control methodologies that emphasize torque control rather than position control. In parallel, we introduce motion planning strategies that minimize jerk and incorporate dynamic constraints, thereby producing trajectories that are both efficient and smooth.This dissertation further enhances adaptability through machine-learning approaches tailored for friction-aware grasping and dynamic manipulation. By constructing a friction coefficient dataset and training regression models, the system can efficiently adjust grasping forces in real-time. Additionally, ongoing research in imitation learning, drawing on human demonstrations and incorporating both visual data and physical models, shows promise in enabling the robot to eventually replicate human-level skill in complex, dynamically varying tasks.Comprehensive evaluations include successful demonstrations of autonomous cooking and friction-aware grasping, as well as ongoing research into dynamic box-receiving tasks. While further validation and optimization remain active areas of study, results so far suggest that, when integrated with torque control, optimized motion planning, and model-based data-driven learning, the proposed proprioceptive manipulator can significantly enhance manipulation speed, stability, energy efficiency, and compliance. These findings provide a strong foundation for more intuitive, versatile, and collaborative robotic systems, better suited to the demands of real-world, human-centered environments.
■590 ▼aSchool code: 0031.
■650 4▼aRobotics
■650 4▼aComputer engineering
■650 4▼aMechanical engineering
■653 ▼aDynamic manipulation
■653 ▼aMachine learning
■653 ▼aOptimal control
■653 ▼aRobotic manipulators
■653 ▼aTrajectory optimization
■690 ▼a0771
■690 ▼a0464
■690 ▼a0548
■71020▼aUniversity of California, Los Angeles▼bMechanical Engineering 0330.
■7730 ▼tDissertations Abstracts International▼g86-06B.
■790 ▼a0031
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17165120▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
