본문

서브메뉴

Safe, High-Performance Motion Planning Under Uncertainty for Autonomous Driving Applications
Safe, High-Performance Motion Planning Under Uncertainty for Autonomous Driving Applicatio...
Safe, High-Performance Motion Planning Under Uncertainty for Autonomous Driving Applications

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260202105511
ISBN  
9798263329372
DDC  
796.72
저자명  
Knaup, Jacob.
서명/저자  
Safe, High-Performance Motion Planning Under Uncertainty for Autonomous Driving Applications
발행사항  
[Sl] : Georgia Institute of Technology, 2024
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2024
형태사항  
168 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-05, Section: A.
주기사항  
Advisor: Tsiotras, Panagiotis.
학위논문주기  
Thesis (Ph.D.)--Georgia Institute of Technology, 2024.
초록/해제  
요약Robots and autonomous vehicles must be able to operate safely and efficiently in the presence of environmental uncertainty, motivating the development of sophisticated motion planning algorithms for handling complex stochastic dynamics. However, the application of existing approaches is limited due to restrictions imposed on the structure of the dynam-ics and conservatism resulting from simplistic characterization of uncertainties. This thesis develops theoretical and experimental results motivated by two challenging robotic appli-cations within the domain of autonomous vehicles: off-road racing and interactive driving. For the first application, autonomous racing, we develop a stochastic model predictive con-trol (SMPC) algorithm for linear time-varying systems subject to unbounded disturbances which efficiently optimizes over the space of affine feedback policies using convex pro-gramming. We show that the proposed SMPC approach provides bounded convergence to a reference trajectory and guarantees polytopic chance constraint satisfaction with at least a specified probability during operation. We verify our approach using numerical and experimental demonstrations on a 1:5-scale autonomous rally racing platform as well as with a full-scale autonomous vehicle, and we demonstrate improved safety and reduced lap-times over a state-of-the-art method. For the second application, we consider interac-tive highway driving and develop a dual SMPC algorithm which employs active learning of unknown parameters of a general nonlinear system using Bayesian estimation. We show that the proposed approach induces probing behaviors to reduce parameter uncertainty, and we develop a novel sampling-based solver utilizing recent results for generative diffusion models to efficiently solve the dual control problem in real time. We validate the proposed dual SMPC approach by utilizing high-fidelity simulations and hardware experiments of an autonomous merge scenario in a congested traffic environment and demonstrate a superior merge success rate over state-of-the-art passive learning approaches.
일반주제명  
Automobile racing
일반주제명  
Control algorithms
일반주제명  
Computer peripherals
일반주제명  
Normal distribution
일반주제명  
Autonomous vehicles
일반주제명  
Roads & highways
일반주제명  
Controllers
일반주제명  
Design
일반주제명  
Stochastic models
일반주제명  
Robotics
일반주제명  
Computer science
일반주제명  
Transportation
기타저자  
Georgia Institute of Technology.
기본자료저록  
Dissertations Abstracts International. 87-05A.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260126s2024        us                              c    eng  d
■001000017360349
■00520260202105511
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798263329372
■035    ▼a(MiAaPQ)AAI32308307
■035    ▼a(MiAaPQ)GeorgiaTech76988
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a796.72
■1001  ▼aKnaup,  Jacob.
■24510▼aSafe,  High-Performance  Motion  Planning  Under  Uncertainty  for  Autonomous  Driving  Applications
■260    ▼a[Sl]▼bGeorgia  Institute  of  Technology▼c2024
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2024
■300    ▼a168  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-05,  Section:  A.
■500    ▼aAdvisor:  Tsiotras,  Panagiotis.
■5021  ▼aThesis  (Ph.D.)--Georgia  Institute  of  Technology,  2024.
■520    ▼aRobots  and  autonomous  vehicles  must  be  able  to  operate  safely  and  efficiently  in  the  presence  of  environmental  uncertainty,  motivating  the  development  of  sophisticated  motion  planning  algorithms  for  handling  complex  stochastic  dynamics.  However,  the  application  of  existing  approaches  is  limited  due  to  restrictions  imposed  on  the  structure  of  the  dynam-ics  and  conservatism  resulting  from  simplistic  characterization  of  uncertainties.  This  thesis  develops  theoretical  and  experimental  results  motivated  by  two  challenging  robotic  appli-cations  within  the  domain  of  autonomous  vehicles:  off-road  racing  and  interactive  driving.  For  the  first  application,  autonomous  racing,  we  develop  a  stochastic  model  predictive  con-trol  (SMPC)  algorithm  for  linear  time-varying  systems  subject  to  unbounded  disturbances  which  efficiently  optimizes  over  the  space  of  affine  feedback  policies  using  convex  pro-gramming.  We  show  that  the  proposed  SMPC  approach  provides  bounded  convergence  to  a  reference  trajectory  and  guarantees  polytopic  chance  constraint  satisfaction  with  at  least  a  specified  probability  during  operation.  We  verify  our  approach  using  numerical  and  experimental  demonstrations  on  a  1:5-scale  autonomous  rally  racing  platform  as  well  as  with  a  full-scale  autonomous  vehicle,  and  we  demonstrate  improved  safety  and  reduced  lap-times  over  a  state-of-the-art  method.  For  the  second  application,  we  consider  interac-tive  highway  driving  and  develop  a  dual  SMPC  algorithm  which  employs  active  learning  of  unknown  parameters  of  a  general  nonlinear  system  using  Bayesian  estimation.  We  show  that  the  proposed  approach  induces  probing  behaviors  to  reduce  parameter  uncertainty,  and  we  develop  a  novel  sampling-based  solver  utilizing  recent  results  for  generative  diffusion  models  to  efficiently  solve  the  dual  control  problem  in  real  time.  We  validate  the  proposed  dual  SMPC  approach  by  utilizing  high-fidelity  simulations  and  hardware  experiments  of  an  autonomous  merge  scenario  in  a  congested  traffic  environment  and  demonstrate  a  superior  merge  success  rate  over  state-of-the-art  passive  learning  approaches.
■590    ▼aSchool  code:  0078.
■650  4▼aAutomobile  racing
■650  4▼aControl  algorithms
■650  4▼aComputer  peripherals
■650  4▼aNormal  distribution
■650  4▼aAutonomous  vehicles
■650  4▼aRoads  &  highways
■650  4▼aControllers
■650  4▼aDesign
■650  4▼aStochastic  models
■650  4▼aRobotics
■650  4▼aComputer  science
■650  4▼aTransportation
■690    ▼a0771
■690    ▼a0389
■690    ▼a0984
■690    ▼a0709
■71020▼aGeorgia  Institute  of  Technology.
■7730  ▼tDissertations  Abstracts  International▼g87-05A.
■790    ▼a0078
■791    ▼aPh.D.
■792    ▼a2024
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17360349▼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
    TF14559 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.