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

Robust Control of Nonlinear Systems ...

Garimella, Gowtham.

FindBook

Google Book

Amazon

博客來

Robust Control of Nonlinear Systems with Applications to Aerial Manipulation and Self Driving Cars.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Robust Control of Nonlinear Systems with Applications to Aerial Manipulation and Self Driving Cars./
作者:	Garimella, Gowtham.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2019,
面頁冊數:	253 p.
附註:	Source: Dissertations Abstracts International, Volume: 81-07, Section: A.
Contained By:	Dissertations Abstracts International81-07A.
標題:	Systems science. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27606835
ISBN:	9781687988003

Robust Control of Nonlinear Systems with Applications to Aerial Manipulation and Self Driving Cars.
Garimella, Gowtham.

Robust Control of Nonlinear Systems with Applications to Aerial Manipulation and Self Driving Cars. - Ann Arbor : ProQuest Dissertations & Theses, 2019 - 253 p.

Source: Dissertations Abstracts International, Volume: 81-07, Section: A.

Thesis (Ph.D.)--The Johns Hopkins University, 2019.

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

This work considers the problem of planning and control of robots in an environment with obstacles and external disturbances. The safety of robots is harder to achieve when planning in such uncertain environments. We describe a robust control scheme that combines three key components: system identification, uncertainty propagation, and trajectory optimization. Using this control scheme we tackle three problems. First, we develop a Nonlinear Model Predictive Controller (NMPC) for articulated rigid bodies and apply it to an aerial manipulation system to grasp object mid-air. Next, we tackle the problem of obstacle avoidance under unknown external disturbances. We propose two approaches, the first approach using adaptive NMPC with open-loop uncertainty propagation and the second approach using Tube NMPC. After that, we introduce dynamic models which use Artificial Neural Networks (ANN) and combine them with NMPC to control a ground vehicle and an aerial manipulation system. Finally, we introduce a software framework for integrating the above algorithms to perform complex tasks. The software framework provides users with the ability to design systems that are robust to control and hardware failures where preventive action is taken before-hand. The framework also allows for safe testing of control and task logic in simulation before evaluating on the real robot. The software framework is applied to an aerial manipulation system to perform a package sorting task, and extensive experiments demonstrate the ability of the system to recover from failures. In addition to robust control, we present two related control problems. The first problem pertains to designing an obstacle avoidance controller for an underactuated system that is Lyapunov stable. We extend a standard gyroscopic obstacle avoidance controller to be applicable to an underactuated system. The second problem addresses the navigation of an Unmanned Ground Vehicle (UGV) on an unstructured terrain. We propose using NMPC combined with a high fidelity physics engine to generate a reference trajectory that is dynamically feasible and accounts for unsafe areas in the terrain.

ISBN: 9781687988003Subjects--Topical Terms:

3168411
Systems science.
Subjects--Index Terms:

Robust control

Robust Control of Nonlinear Systems with Applications to Aerial Manipulation and Self Driving Cars.
LDR:03511nmm a2200433 4500 001 2272921
005 20201105110259.5
008 220629s2019 ||||||||||||||||| ||eng d
020 $a 9781687988003
035 $a (MiAaPQ)AAI27606835
035 $a (MiAaPQ)0098vireo4405Garimella
035 $a AAI27606835
040 $a MiAaPQ $c MiAaPQ
100 1 $a Garimella, Gowtham. $3 3550346
245 1 0 $a Robust Control of Nonlinear Systems with Applications to Aerial Manipulation and Self Driving Cars.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2019
300 $a 253 p.
500 $a Source: Dissertations Abstracts International, Volume: 81-07, Section: A.
500 $a Advisor: Kobilarov, Marin.
502 $a Thesis (Ph.D.)--The Johns Hopkins University, 2019.
506 $a This item must not be sold to any third party vendors.
520 $a This work considers the problem of planning and control of robots in an environment with obstacles and external disturbances. The safety of robots is harder to achieve when planning in such uncertain environments. We describe a robust control scheme that combines three key components: system identification, uncertainty propagation, and trajectory optimization. Using this control scheme we tackle three problems. First, we develop a Nonlinear Model Predictive Controller (NMPC) for articulated rigid bodies and apply it to an aerial manipulation system to grasp object mid-air. Next, we tackle the problem of obstacle avoidance under unknown external disturbances. We propose two approaches, the first approach using adaptive NMPC with open-loop uncertainty propagation and the second approach using Tube NMPC. After that, we introduce dynamic models which use Artificial Neural Networks (ANN) and combine them with NMPC to control a ground vehicle and an aerial manipulation system. Finally, we introduce a software framework for integrating the above algorithms to perform complex tasks. The software framework provides users with the ability to design systems that are robust to control and hardware failures where preventive action is taken before-hand. The framework also allows for safe testing of control and task logic in simulation before evaluating on the real robot. The software framework is applied to an aerial manipulation system to perform a package sorting task, and extensive experiments demonstrate the ability of the system to recover from failures. In addition to robust control, we present two related control problems. The first problem pertains to designing an obstacle avoidance controller for an underactuated system that is Lyapunov stable. We extend a standard gyroscopic obstacle avoidance controller to be applicable to an underactuated system. The second problem addresses the navigation of an Unmanned Ground Vehicle (UGV) on an unstructured terrain. We propose using NMPC combined with a high fidelity physics engine to generate a reference trajectory that is dynamically feasible and accounts for unsafe areas in the terrain.
590 $a School code: 0098.
650 4 $a Systems science. $3 3168411
650 4 $a Transportation. $3 555912
650 4 $a Robotics. $3 519753
653 $a Robust control
653 $a Tube nmpc
653 $a Stochastic systems
653 $a Planning
653 $a Controls
653 $a Uncertainty
653 $a Aerial manipulation
653 $a Neural networks
653 $a Gyroscopic avoidance
690 $a 0771
690 $a 0790
690 $a 0709
710 2 $a The Johns Hopkins University. $b Mechanical Engineering. $3 3175788
773 0 $t Dissertations Abstracts International $g 81-07A.
790 $a 0098
791 $a Ph.D.
792 $a 2019
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27606835