본문

서브메뉴

A Safe and Robust Multi-Agent Motion Planning Framework for Urban Air Mobility
A Safe and Robust Multi-Agent Motion Planning Framework for Urban Air Mobility
A Safe and Robust Multi-Agent Motion Planning Framework for Urban Air Mobility

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260202105329
ISBN  
9798263325114
DDC  
796.51
저자명  
Netter, Joshua.
서명/저자  
A Safe and Robust Multi-Agent Motion Planning Framework for Urban Air Mobility
발행사항  
[Sl] : Georgia Institute of Technology, 2025
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2025
형태사항  
117 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-05, Section: B.
주기사항  
Advisor: Vamvoudakis, Kyriakos G.
학위논문주기  
Thesis (Ph.D.)--Georgia Institute of Technology, 2025.
초록/해제  
요약In this dissertation, we consider numerous challenges that face the development of urban air mobility (UAM). We formulate the UAM problem, and propose a motion planning framework for a number of cooperating "player agents" in the environment. We then propose a model-free method for each player agent to learn their optimal control. We continue by considering the potential behavior of other agents in the environment, and propose algorithms for navigating around cooperative, independent, and adversarial agents. We formulate a distributed path planning algorithm for cooperative agents to safely reach their destinations while minimizing the chance of deadlock. We continue by constructing a cognitive hierarchy to identify potential strategies in independent agents and identify these strategies using distributed Gaussian process classification to ensure player agents can avoid these independent agents. Last, we demonstrate an algorithm to predict if any of these independent agents are adversarial agents pursuing our player agents and formulate a motion planning approach to avoid these potential adversaries. The efficacy of this complete motion planning algorithm is shown in numerous simulations.We also consider the challenges of modeling the constraints present on a UAV system. Given some amount of "quantities of interest" (QoIs), we propose an efficient method to learn a binary "go/no-go" classifier by selecting informative training points via a soft actor/critic (SAC) framework to minimize simulation runtime during the system's design. We continue on to use the learned actor-critic to augment target system inputs in order to ensure safety online. We also incorporate this actor-critic with our previous method of model-free optimal control to drive our learned optimal control away from constraints online. Finally, we use the learned actor-critic to make online observations of the system, and predict if the system is faulty or unsafe due to hardware failure or an attack. The efficacy of both our informed sampling approach to train a binary classifier, as well as actor-critic framework's potential online uses, are shown in simulations.
일반주제명  
Trails
일반주제명  
Robots
일반주제명  
Design
일반주제명  
Drones
일반주제명  
Rationality
일반주제명  
Planning
일반주제명  
Aerial surveying
일반주제명  
Robotics
기타저자  
Georgia Institute of Technology.
기본자료저록  
Dissertations Abstracts International. 87-05B.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260126s2025        us                              c    eng  d
■001000017360256
■00520260202105329
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798263325114
■035    ▼a(MiAaPQ)AAI32307925
■035    ▼a(MiAaPQ)GeorgiaTech78703
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a796.51
■1001  ▼aNetter,  Joshua.
■24512▼aA  Safe  and  Robust  Multi-Agent  Motion  Planning  Framework  for  Urban  Air  Mobility
■260    ▼a[Sl]▼bGeorgia  Institute  of  Technology▼c2025
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2025
■300    ▼a117  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-05,  Section:  B.
■500    ▼aAdvisor:  Vamvoudakis,  Kyriakos  G.
■5021  ▼aThesis  (Ph.D.)--Georgia  Institute  of  Technology,  2025.
■520    ▼aIn  this  dissertation,  we  consider  numerous  challenges  that  face  the  development  of  urban  air  mobility  (UAM).  We  formulate  the  UAM  problem,  and  propose  a  motion  planning  framework  for  a  number  of  cooperating  "player  agents"  in  the  environment.  We  then  propose  a  model-free  method  for  each  player  agent  to  learn  their  optimal  control.  We  continue  by  considering  the  potential  behavior  of  other  agents  in  the  environment,  and  propose  algorithms  for  navigating  around  cooperative,  independent,  and  adversarial  agents.  We  formulate  a  distributed  path  planning  algorithm  for  cooperative  agents  to  safely  reach  their  destinations  while  minimizing  the  chance  of  deadlock.  We  continue  by  constructing  a  cognitive  hierarchy  to  identify  potential  strategies  in  independent  agents  and  identify  these  strategies  using  distributed  Gaussian  process  classification  to  ensure  player  agents  can  avoid  these  independent  agents.  Last,  we  demonstrate  an  algorithm  to  predict  if  any  of  these  independent  agents  are  adversarial  agents  pursuing  our  player  agents  and  formulate  a  motion  planning  approach  to  avoid  these  potential  adversaries.  The  efficacy  of  this  complete  motion  planning  algorithm  is  shown  in  numerous  simulations.We  also  consider  the  challenges  of  modeling  the  constraints  present  on  a  UAV  system.  Given  some  amount  of  "quantities  of  interest"  (QoIs),  we  propose  an  efficient  method  to  learn  a  binary  "go/no-go"  classifier  by  selecting  informative  training  points  via  a  soft  actor/critic  (SAC)  framework  to  minimize  simulation  runtime  during  the  system's  design.  We  continue  on  to  use  the  learned  actor-critic  to  augment  target  system  inputs  in  order  to  ensure  safety  online.  We  also  incorporate  this  actor-critic  with  our  previous  method  of  model-free  optimal  control  to  drive  our  learned  optimal  control  away  from  constraints  online.  Finally,  we  use  the  learned  actor-critic  to  make  online  observations  of  the  system,  and  predict  if  the  system  is  faulty  or  unsafe  due  to  hardware  failure  or  an  attack.  The  efficacy  of  both  our  informed  sampling  approach  to  train  a  binary  classifier,  as  well  as  actor-critic  framework's  potential  online  uses,  are  shown  in  simulations.
■590    ▼aSchool  code:  0078.
■650  4▼aTrails
■650  4▼aRobots
■650  4▼aDesign
■650  4▼aDrones
■650  4▼aRationality
■650  4▼aPlanning
■650  4▼aAerial  surveying
■650  4▼aRobotics
■690    ▼a0389
■690    ▼a0771
■71020▼aGeorgia  Institute  of  Technology.
■7730  ▼tDissertations  Abstracts  International▼g87-05B.
■790    ▼a0078
■791    ▼aPh.D.
■792    ▼a2025
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17360256▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

미리보기

내보내기

chatGPT토론

Ai 추천 관련 도서


    신착도서 더보기
    최근 3년간 통계입니다.

    소장정보

    • 예약
    • 소재불명신고
    • 나의폴더
    • 우선정리요청
    • 비도서대출신청
    • 야간 도서대출신청
    소장자료
    등록번호 청구기호 소장처 대출가능여부 대출정보
    TF18581 전자도서 대출가능 마이폴더 부재도서신고 비도서대출신청 야간 도서대출신청

    * 대출중인 자료에 한하여 예약이 가능합니다. 예약을 원하시면 예약버튼을 클릭하십시오.

    해당 도서를 다른 이용자가 함께 대출한 도서

    관련 인기도서

    로그인 후 이용 가능합니다.