Xudong Yao

Publication Date

1996

Document Type

Dissertation/Thesis

First Advisor

Bobis, James P.

Degree Name

M.S. (Master of Science)

Legacy Department

Department of Electrical Engineering

LCSH

Electronic controllers; Neural networks (Computer science); Nonlinear systems; Fuzzy systems

Abstract

Physical systems are inherently nonlinear. Nonlinear system can be described by nonlinear differential equations. Some common nonlinear system behaviors are multiple equilibrium points, limit cycles, bifurcation, chaos. For nonlinear control, the research task involved constructing a controller so that the closed loop system meets the desired characteristics. This paper shows the construction process for a neural-fuzzy robust adaptive position controller using MATLAB. For this project's controller, the fuzzy approach is used in order to convert human knowledge into control knowledge. The fuzzy membership functions and rules were implemented with both a backpropagation feedforward neural network and a logical network. A novel devised adaptive algorithm is used to adjust the centroid parameter. To demonstrate the robustness (which is the ability to control different nonlinear models) of this controller, two first order and two second order nonlinear plants, a chaotic system, and a pendulum model are used to implement the simulation. The results show that the controller cancels the nonlinear terms in the system and causes the system to converge to step inputs.

Comments

Includes bibliographical references (leaf [72])

Extent

viii, 84 pages

Language

eng

Publisher

Northern Illinois University

Rights Statement

In Copyright

Rights Statement 2

NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.

Media Type

Text

Download

Share

COinS