Doctoral Dissertations

Author

Hao Xu

Abstract

"Networked Control System (NCS) is a recent topic of research wherein the feedback control loops are closed through a real-time communication network. Many design challenges surface in such systems due to network imperfections such as random delays, packet losses, quantization effects and so on. Since existing control techniques are unsuitable for such systems, in this dissertation, a suite of novel stochastic optimal adaptive design methodologies is undertaken for both linear and nonlinear NCS in presence of uncertain system dynamics and unknown network imperfections such as network-induced delays and packet losses. The design is introduced in five papers.

In Paper 1, a stochastic optimal adaptive control design is developed for unknown linear NCS with uncertain system dynamics and unknown network imperfections. A value function is adjusted forward-in-time and online, and a novel update law is proposed for tuning value function estimator parameters. Additionally, by using estimated value function, optimal adaptive control law is derived based on adaptive dynamic programming technique. Subsequently, this design methodology is extended to solve stochastic optimal strategies of linear NCS zero-sum games in Paper 2.

Since most systems are inherently nonlinear, a novel stochastic optimal adaptive control scheme is then developed in Paper 3 for nonlinear NCS with unknown network imperfections. On the other hand, in Paper 4, the network protocol behavior (e.g. TCP and UDP) are considered and optimal adaptive control design is revisited using output feedback for linear NCS. Finally, Paper 5 explores a co-design framework where both the controller and network scheduling protocol designs are addressed jointly so that proposed scheme can be implemented into next generation Cyber Physical Systems"--Abstract, page iv.

Advisor(s)

Sarangapani, Jagannathan, 1965-

Committee Member(s)

Zheng, Y. Rosa
Balakrishnan, S. N.
Madria, Sanjay Kumar
Zawodniok, Maciej Jan, 1975

Department(s)

Electrical and Computer Engineering

Degree Name

Ph. D. in Electrical Engineering

Sponsor(s)

Missouri University of Science and Technology. Intelligent Systems Center
National Science Foundation (U.S.)

Publisher

Missouri University of Science and Technology

Publication Date

Summer 2012

Journal article titles appearing in thesis/dissertation

  • Stochastic optimal design for unknown linear discrete-time system zero-sum games under communication constraints
  • Stochastic optimal controller design for uncertain nonlinear networked control system via neuro dynamic programming
  • Stochastic optimal design for unknown networked control system with communication network protocols
  • Optimal control design for NCS under UDP
  • Cross layer approach to the novel distributed scheduling protocol and event-triggered controller design for cyber physical systems

Pagination

xv, 261 pages

Note about bibliography

Includes bibliographical references.

Rights

© 2012 Hao Xu, All rights reserved.

Document Type

Dissertation - Open Access

File Type

text

Language

English

Library of Congress Subject Headings

Computer network protocols
Embedded computer systems
Stochastic control theory
Wireless sensor networks -- Design

Thesis Number

T 10040

Print OCLC #

815767805

Electronic OCLC #

808442049

Share

 
COinS