viii PREFACE
To help meet thefirst objective, we provide the reader a broadflavor of what
classical control theory involves, and we present in some depth the mechanics of
implementing classical control techniques. It is not our intent to cover classical
methods in great detail as much as to provide the reader with afirm understand-
ing of the principles that govern system behavior and control. As an outcome of
this presentation, the type of information needed to implement classical control
techniques and some of the limitations of classical control techniques should
become obvious to the reader.
Our second objective is to present sufficient background in both fuzzy and
neural control so that further studies can be pursued in advanced soft comput-
ing methodologies. The emphasis in this presentation is to demonstrate the
ease with which system control can be achieved in the absence of an analytical
mathematical model. The benefits of a model-free methodology in comparison
with a model-based methodology for control are made clear. Again, it is our in-
tent to bring to the reader the fundamental mechanics of both fuzzy and neural
control technologies and to demonstrate clearly how such methodologies can be
implemented for nonlinear system control.
This text,A First Course in Fuzzy and Neural Control,is intended to address
all the material needed to motivate students towards further studies in soft
computing. Our intent is not to overwhelm students with unnecessary material,
either from a mathematical or engineering perspective, but to provide balance
between the mathematics and engineering aspects of fuzzy and neural network-
based approaches. In fact, we strongly recommend that students acquire the
mathematical foundations and knowledge of standard control systems before
taking a course in soft computing methods.
Chapter 1 provides the fundamental ideas of control theory through simple
examples. Our goal is to show the consequences of systems that either do or
do not have feedback, and to provide insights into controller design concepts.
From these examples it should become clear that systems can be controlled if
they exhibit the two properties ofcontrollabilityandobservability.
Chapter 2 provides a background of classical control methodologies, in-
cluding state-variable approaches, thatform the basis for control systems de-
sign. We discuss state-variable and output feedback as the primary moti-
vation for designing controllers via pole-placement for systems that are in-
herently unstable. We extend these classical control concepts to the design
of conventionalProportional-Integral (PI),Proportional-Derivative(PD), and
Proportional-Integral-Derivative (PID) controllers. Chapter 2 includes a dis-
cussion of stability and classical methods of determining stability of nonlinear
systems.
Chapter 3 introduces mathematical notions used in linguistic rule-based con-
trol. In this context, several basic examples are discussed that lay the mathe-
matical foundations of fuzzy set theory. We introduce linguistic rules ñ methods
for inferencing based on the mathematical theory of fuzzy sets. This chapter
emphasizes the logical aspects of reasoning needed for intelligent control and
decision support systems.
In Chapter 4, we present an introduction to fuzzy control, describing the