서브메뉴
검색
Human Supervision of Multi-Robot Systems
Human Supervision of Multi-Robot Systems
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202105231
- ISBN
- 9798291567258
- DDC
- 620
- 저자명
- Gilbert, Alia L.
- 서명/저자
- Human Supervision of Multi-Robot Systems
- 발행사항
- [Sl] : University of Michigan, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 85 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-03, Section: B.
- 주기사항
- Advisor: Panagou, Dimitra.
- 학위논문주기
- Thesis (Ph.D.)--University of Michigan, 2025.
- 초록/해제
- 요약As autonomous systems grow more capable and are increasingly deployed across diverse domains, the question of how humans supervise these systems has become increasingly important. Supervision today is not limited to overseeing actions or responding to alerts. It involves interpreting system behavior, managing uncertainty, and intervening when autonomy falls short. This dissertation examines how different system and interface designs can better support human supervision across a range of configurations in which the roles of human and machine are dynamic and interdependent.Rather than concentrating on systems with either fully manual or fully autonomous control, this work focuses on the middle of the autonomy spectrum, where control is shared and effective supervision depends on how well information is communicated between agents. Across four case studies, it investigates how system design can support human engagement in ways that are flexible, interpretable, and robust to context changes and system limitations.Chapter 2 presents a control framework for aerial robots that maintain formation around a human leader whose trajectory is not explicitly communicated. The system uses set-theoretic methods to maintain safety and connectivity under uncertainty, demonstrating how autonomous systems can remain sensitive to human behavior even without direct commands. Chapter 3 introduces a mixed reality interface that allows users to supervise ground robots remotely by issuing goals and viewing spatialized sensor feedback through a head-mounted display. The system was demonstrated in live outreach settings and contributes an opensource tool for immersive human-robot supervision. Chapter 4 evaluates whether passive haptic access to an automated vehicle's control signal can improve human performance during supervisory handovers. In a controlled experiment, participants used a shared joystick to feel the vehicle's actions before taking over, and results showed improved tracking accuracy and better secondary task performance. Chapter 5 examines how interface modality affects human intervention during failure recovery. In particular, it compares mouse and keyboard input for helping a ground robot escape a local minimum, showing that even subtle interface design choices influence intervention timing and success rates.Together, these studies contribute empirical findings, system prototypes, and design insights that inform the development of supervisory systems for multi-robot teams. The dissertation emphasizes that effective supervision depends not only on when control shifts, but also on how humans remain perceptually and functionally engaged across different roles and task complexities. By studying a range of interaction modes and control configurations, this work highlights strategies for designing human-autonomy systems that are more resilient, communicative, and attuned to shared responsibilities.
- 일반주제명
- Engineering
- 일반주제명
- Robotics
- 키워드
- Aerial robots
- 키워드
- Human leader
- 기타저자
- University of Michigan Robotics
- 기본자료저록
- Dissertations Abstracts International. 87-03B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017359883
■00520260202105231
■006m o d
■007cr#unu||||||||
■020 ▼a9798291567258
■035 ▼a(MiAaPQ)AAI32271889
■035 ▼a(MiAaPQ)umichrackham006298
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a620
■1001 ▼aGilbert, Alia L.
■24510▼aHuman Supervision of Multi-Robot Systems
■260 ▼a[Sl]▼bUniversity of Michigan▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a85 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-03, Section: B.
■500 ▼aAdvisor: Panagou, Dimitra.
■5021 ▼aThesis (Ph.D.)--University of Michigan, 2025.
■520 ▼aAs autonomous systems grow more capable and are increasingly deployed across diverse domains, the question of how humans supervise these systems has become increasingly important. Supervision today is not limited to overseeing actions or responding to alerts. It involves interpreting system behavior, managing uncertainty, and intervening when autonomy falls short. This dissertation examines how different system and interface designs can better support human supervision across a range of configurations in which the roles of human and machine are dynamic and interdependent.Rather than concentrating on systems with either fully manual or fully autonomous control, this work focuses on the middle of the autonomy spectrum, where control is shared and effective supervision depends on how well information is communicated between agents. Across four case studies, it investigates how system design can support human engagement in ways that are flexible, interpretable, and robust to context changes and system limitations.Chapter 2 presents a control framework for aerial robots that maintain formation around a human leader whose trajectory is not explicitly communicated. The system uses set-theoretic methods to maintain safety and connectivity under uncertainty, demonstrating how autonomous systems can remain sensitive to human behavior even without direct commands. Chapter 3 introduces a mixed reality interface that allows users to supervise ground robots remotely by issuing goals and viewing spatialized sensor feedback through a head-mounted display. The system was demonstrated in live outreach settings and contributes an opensource tool for immersive human-robot supervision. Chapter 4 evaluates whether passive haptic access to an automated vehicle's control signal can improve human performance during supervisory handovers. In a controlled experiment, participants used a shared joystick to feel the vehicle's actions before taking over, and results showed improved tracking accuracy and better secondary task performance. Chapter 5 examines how interface modality affects human intervention during failure recovery. In particular, it compares mouse and keyboard input for helping a ground robot escape a local minimum, showing that even subtle interface design choices influence intervention timing and success rates.Together, these studies contribute empirical findings, system prototypes, and design insights that inform the development of supervisory systems for multi-robot teams. The dissertation emphasizes that effective supervision depends not only on when control shifts, but also on how humans remain perceptually and functionally engaged across different roles and task complexities. By studying a range of interaction modes and control configurations, this work highlights strategies for designing human-autonomy systems that are more resilient, communicative, and attuned to shared responsibilities.
■590 ▼aSchool code: 0127.
■650 4▼aEngineering
■650 4▼aRobotics
■653 ▼aHuman-robot interaction
■653 ▼aAerial robots
■653 ▼aHuman leader
■653 ▼aMulti-robot teams
■690 ▼a0771
■690 ▼a0537
■690 ▼a0800
■71020▼aUniversity of Michigan▼bRobotics.
■7730 ▼tDissertations Abstracts International▼g87-03B.
■790 ▼a0127
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17359883▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
