東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

FindBook

Google Book

Amazon

博客來

Differentiable and Bilevel Optimization for Control in Robotics.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Differentiable and Bilevel Optimization for Control in Robotics./
作者:	Landry, Benoit.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2021,
面頁冊數:	112 p.
附註:	Source: Dissertations Abstracts International, Volume: 83-05, Section: B.
Contained By:	Dissertations Abstracts International83-05B.
標題:	Friction. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28688347
ISBN:	9798544204077

Differentiable and Bilevel Optimization for Control in Robotics.
Landry, Benoit.

Differentiable and Bilevel Optimization for Control in Robotics. - Ann Arbor : ProQuest Dissertations & Theses, 2021 - 112 p.

Source: Dissertations Abstracts International, Volume: 83-05, Section: B.

Thesis (Ph.D.)--Stanford University, 2021.

This item must not be sold to any third party vendors.

Over the next few decades, much of the western world will have to wrestle with the consequences of an aging population. Indeed, the declining population growth rates in the West [1] are all but guaranteed to lead to situations where cohorts of working age are confronted with the monumental task of caring for significantly bigger, older ones. This situation will call for all sorts of creative solutions, and robotics is likely to play an important role in this. Even though we have already seen robotics starting to take its place in a few key sectors such as transportation and warehousing, it's important to identify the current limitations of these systems. Indeed, the robotics community has made great strides in developing highly effective algorithms for capabilities like obstacle avoidance, which is the fundamental building block of many of the existing applications of robotics to unstructured environments. It is however obvious that the performance of these systems quickly degrades as soon as they are asked to perform more complex tasks, such as ones that involve making and breaking contact with their environment. In this dissertation, we take inspiration from one of the areas we believe has fundamentally enabled the most successful of these deployments over the last few years: numerical optimization. More specifically, we investigate an up-and-coming class of mathematical programs, bilevel optimization, and how it can be leveraged to tackle the most pressing algorithmic challenges of control in robotics.Bilevel optimization, where two mathematical programs are nested into one another, has recently gained renewed attention thanks to newfound abilities to solve these problems efficiently. In this dissertation, we give an overview of our work on bilevel optimization, and our progress on leveraging this class of problems to move us closer to computationally tractable control of nonlinear systems. Specifically, we demonstrate how it is possible to design novel solution methods that utilize advances in automatic differentiation while retaining the benefits of state of the art constrained nonlinear optimization solvers. We also demonstrate how particularly challenging problems of nonlinear control such as planning through contact, adversarial learning of value functions, and Lyapunov synthesis can all surprisingly be tackled by explicitly addressing them as bilevel optimization problems.

ISBN: 9798544204077Subjects--Topical Terms:

650299
Friction.

Differentiable and Bilevel Optimization for Control in Robotics.
LDR:03450nmm a2200325 4500 001 2344867
005 20220531062201.5
008 241004s2021 ||||||||||||||||| ||eng d
020 $a 9798544204077
035 $a (MiAaPQ)AAI28688347
035 $a (MiAaPQ)STANFORDbw199zy3697
035 $a AAI28688347
040 $a MiAaPQ $c MiAaPQ
100 1 $a Landry, Benoit. $3 3683696
245 1 0 $a Differentiable and Bilevel Optimization for Control in Robotics.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2021
300 $a 112 p.
500 $a Source: Dissertations Abstracts International, Volume: 83-05, Section: B.
500 $a Advisor: Pavone, Marco.
502 $a Thesis (Ph.D.)--Stanford University, 2021.
506 $a This item must not be sold to any third party vendors.
520 $a Over the next few decades, much of the western world will have to wrestle with the consequences of an aging population. Indeed, the declining population growth rates in the West [1] are all but guaranteed to lead to situations where cohorts of working age are confronted with the monumental task of caring for significantly bigger, older ones. This situation will call for all sorts of creative solutions, and robotics is likely to play an important role in this. Even though we have already seen robotics starting to take its place in a few key sectors such as transportation and warehousing, it's important to identify the current limitations of these systems. Indeed, the robotics community has made great strides in developing highly effective algorithms for capabilities like obstacle avoidance, which is the fundamental building block of many of the existing applications of robotics to unstructured environments. It is however obvious that the performance of these systems quickly degrades as soon as they are asked to perform more complex tasks, such as ones that involve making and breaking contact with their environment. In this dissertation, we take inspiration from one of the areas we believe has fundamentally enabled the most successful of these deployments over the last few years: numerical optimization. More specifically, we investigate an up-and-coming class of mathematical programs, bilevel optimization, and how it can be leveraged to tackle the most pressing algorithmic challenges of control in robotics.Bilevel optimization, where two mathematical programs are nested into one another, has recently gained renewed attention thanks to newfound abilities to solve these problems efficiently. In this dissertation, we give an overview of our work on bilevel optimization, and our progress on leveraging this class of problems to move us closer to computationally tractable control of nonlinear systems. Specifically, we demonstrate how it is possible to design novel solution methods that utilize advances in automatic differentiation while retaining the benefits of state of the art constrained nonlinear optimization solvers. We also demonstrate how particularly challenging problems of nonlinear control such as planning through contact, adversarial learning of value functions, and Lyapunov synthesis can all surprisingly be tackled by explicitly addressing them as bilevel optimization problems.
590 $a School code: 0212.
650 4 $a Friction. $3 650299
650 4 $a Robots. $3 529507
650 4 $a Algorithms. $3 536374
650 4 $a Optimization. $3 891104
650 4 $a Neural networks. $3 677449
650 4 $a Parameter estimation. $3 567557
650 4 $a Robotics. $3 519753
650 4 $a Artificial intelligence. $3 516317
650 4 $a Computer science. $3 523869
690 $a 0771
690 $a 0800
690 $a 0984
710 2 $a Stanford University. $3 754827
773 0 $t Dissertations Abstracts International $g 83-05B.
790 $a 0212
791 $a Ph.D.
792 $a 2021
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28688347