KafrelSheikh University

Faculty of Engineering
Electrical Engineering Department K ‫جام ـ عة ك ـفر ال ش ـ ـ ـ ـ يخ‬
afrelsheikh University

Intelligent Control

Systems
Fourth Year

Prepared by
Dr. Wessam M. Fikry
2024 - 2025
Chapter: 1

Introduction to Intelligent Control

Systems

C ontrol systems are highly required to have better dynamic performance,

yet are expected to cope with more complex, uncertain, and highly
nonlinear dynamic processes. This in turn limits the use of conventional
control approaches based on differential and difference equations. This is
because the success of conventional control design techniques is based mainly
upon knowledge of the process mathematical model. Moreover, while
conventional control theory typically considers nominal operating conditions
and small variation from these conditions, industrial Practitioners widely agree
that a design that delivers excellent performance under nominal conditions
may produce poor results when operating conditions vary. This has given rise
to the development of intelligent control systems.

Intelligent control achieves automation via the emulation of biological

intelligence. It either seeks to replace a human who performs a control task
(e.g., a chemical process operator) or it borrows ideas from how biological
systems solve problems and apply them to the solution of control problems.

It covers a broad range of topics, such as artificial intelligence, operations

research, fuzzy logic, neural networks, and fuzzy neural networks as well as
genetic algorithms.

1
Chapter One
Introduction to Intelligent Control Systems
Among these topics, fuzzy logic controller (FLC), neural networks
(NNS), fuzzy neural networks (FNNs), and genetic algorithms (GAs) have
played an important role in the development of intelligent control systems
because of their suitability for this work. Many researchers have studied
methodologies for combining these techniques with control theory to build
high-performance controllers for complex systems involving imprecisely
known and nonlinear dynamics.

1.1 Conventional Control System Design

The design of a conventional controller that can alter or modify the
behavior and response of an unknown plant to meet certain performance
requirements can be a tedious and challenging problem in many control
applications. By plant, we mean any process characterized by a certain number
of inputs u and outputs y, as shown in Fig. 1.1. The plant inputs u is processed
to produce several plant outputs y that represent the measured output response
of the plant. The control design task is to choose the input u so that the output
response y(t) satisfies certain given performance requirements.

Fig.1.1 Plant representation.

Because the plant process is usually complex, i.e., it may consist of various
mechanical, electronic, hydraulic parts, etc., the appropriate choice of u is in
Chapter One
Introduction to Intelligent Control Systems
general not straightforward. Thus, there are three basic steps that must be
followed by most control engineering in choosing the input u, which is known
as the controller design steps. Basically, these are modelling, controller
design, and implementation as shown in Fig. 1.2 and are explained below.

Fig.1.2 Control system design step

1.1.1 Mathematical Modelling

Chapter One
Introduction to Intelligent Control Systems
The task of the control engineer in this step is to understand the
processing mechanism of the plant, which takes a given input signal u (t) and
produces the output response y (t), to the point that it can be describe in the
form of some mathematical equations. These equations constitute the
mathematical model of the plant.

Basically, there are only a few ways to generate the model. The first
principles of physics law can be used to write down a model. Another way is
to perform “system identification” via the use of real plant data to produce
a model of the system. Sometimes a combined approach is used where we
use physics to write down a general differential equation that we believe
represents the plant behavior, and then we perform experiments on the plant
to determine certain model parameters or functions.

Often, more than one mathematical model is produced. A “truth model”

is one that is developed to be as accurate as possible so that it can be used in
simulation-based evaluations of control systems.

Linear models such as given below have been used extensively in the past
and the control theory for linear systems is quite mature.
x= Ax + Bu
y = cx + DU
In this case u is the m-dimensional input; x is the n-dimensional state, y is the
p dimensional output; and A, B, C, and D are matrices of appropriate
dimension. Such models, or transfer functions:

G(s) = C(sI − A)−1B + D

Chapter One
Introduction to Intelligent Control Systems
where s is the Laplace variable, are appropriate for use with frequency domain
design techniques (e.g., Bode plots and Nyquist plots), the root-locus method,
state-space methods, and so on. Sometimes it is assumed that the parameters
of the linear model are constant but unknown or can be perturbed from their
nominal values (then techniques for “robust control” or adaptive control are
developed).

Much of the current focus in control is on the development of controllers using

nonlinear models of the plant of the form:
x= f (x,u)
y = g(x,u)
where the variables are defined as for the linear model and f and g are
nonlinear functions of their arguments.

Of particular interest with the above nonlinear models is the case where f
and g are not completely known, and subsequent research focuses on robust
control of nonlinear systems.

Discrete time versions of the above models are also used, and stochastic
effects are often considered via the addition of a random input or other
stochastic effects. Under certain assumptions the nonlinear model in Equation
(1.2) can be linearized to obtain a linear one.

In fact, there are certain properties of the plant that the control engineer
often seeks to identify early in the design process. For instance,
 The stability of the plant may be analyzed (e.g., to see if certain variables
remain bounded).
Chapter One
Introduction to Intelligent Control Systems
 The effects of certain nonlinearities are also studied.

 The engineer may want to determine if the plant is “controllable” to see,

for example, if the control inputs will be able to properly affect the plant,

 The engineer may want to determine if the plant is “observable” to see,

for example, if the chosen sensors will allow the controller to observe the
critical plant behavior so that it can be compensated for, or,

 If it is “non-minimum phase.”

These properties will have a fundamental impact on our ability to design

effective controllers for the system. In addition, the engineer will try to make
a general assessment of how the plant behaves under various conditions, how
the plant dynamics may change over time, and what random effects are
present. Overall, this analysis of the plant’s behavior gives the control
engineer a fundamental understanding of the plant dynamics. This will be very
valuable when it comes time to synthesize a controller.

However, it must be understood, the complexity of most physical plants

makes the development of such an exact model unwarranted or even
impossible. But even if the exact plant model becomes available, its dimension
is likely to be infinite, and its description is nonlinear or time varying.

The mathematical model is an abstraction and hence cannot perfectly

represent all possible dynamics of any physical process (e.g., certain noise
characteristics or failure conditions). This is not to say that we cannot
produce models that are “accurate enough” to closely represent the behavior
of a physical system. Usually, control engineers keep in mind that for control
Chapter One
Introduction to Intelligent Control Systems
design they only need to use a model that is accurate enough to be able to
design a controller that will work. Then, they often also need a very accurate
model to test the controller in simulation (e.g., the truth model) before it is
tested in an experimental setting. Hence, some of the approaches are often
used to obtain a simplified model. They are:
(1) Linearization around operating points

In this approach, the plant is approximated by a linear model that is valid

around a given operating point. Different operating points may lead to several
different linear models that are used as plant models. Linearization is achieved
by using Taylor's series expansion and approximation, fitting experimental
data to a linear model, etc.

(2) Model order reduction techniques

In this approach, small effects, and phenomena outside the frequency range of
interest are neglected leading to a lower order and simpler plant model.

It is convenient to emphasize that the reason that often-simpler models

are developed is that the synthesis techniques for controllers often require that
the model of the plant satisfy certain assumptions (e.g., linearity) or these
methods generally cannot be used.

1.1.2 Controller Design

Once a model of the plant is available, one can proceed with the
controller design. In fact, the controller is designed to meet the performance
requirements for the plant model. These requirements can involve the
following factors:
Chapter One
Introduction to Intelligent Control Systems
 Disturbance rejection properties
 Insensitivity to plant parameter variation
 Stability
 Rise-time
 Overshoot
 Settling time

While these factors are used to characterize the technical conditions that
indicate whether a control system is performing properly, there are other
issues that must be considered that are often of equal or greater importance.
These include the following:
 Cost: How much money will it take to implement the controller, or how
much time will it take to develop the controller?
 Computational complexity: How much processor power and memory will
it take to implement the controller?
 Manufacturability: Does your controller have any extraordinary
requirements about manufacturing the hardware that is to implement it?

Basically, conventional control has provided numerous methods for

constructing controllers for dynamic systems. Some of these are listed below:
► Proportional-integral-derivative (PID) control: Over 90% of the
controllers in operation today are PID controllers (or at least some form of
PID controller like a P or PI controller). This approach is often viewed as
simple, reliable, and easy to understand.

► Classical control: Lead-lag compensation, Bode and Nyquist methods,

root-locus design, and so on.
Chapter One
Introduction to Intelligent Control Systems
► State-space methods: State feedback, observers, and so on.

► Optimal control: Linear quadratic regulator, use of Pontryagin’s

minimum principle or dynamic programming, and so on.

► Robust control: H2 or H∞ methods, quantitative feedback theory, loop

shaping, and so on.

► Nonlinear methods: Feedback linearization, Lyapunov redesign, sliding

mode control, backstepping, and so on.

► Adaptive control: Model reference adaptive control, self-tuning

regulators, nonlinear adaptive control, and so on.

► Stochastic control: Minimum variance control, linear quadratic gaussian

(LQG) control, stochastic adaptive control, and so on.

► Discrete event systems: Petri nets, supervisory control, infinitesimal

perturbation analysis, and so on.

1.1.3 Implementation

In this step, a controller designed in Step 2, which is shown to meet the

performance requirements for the plant model, is ready to be applied to the
unknown plant. The implementation can be done using a digital computer.
Issues, such as the type of computer available, the type of interface devices
between the computer and the plant, software tools, etc., need to be considered
a priori.

Computer speed and accuracy limitations may put constraints on the

complexity of the controller that may force the control engineer to go back to
Chapter One
Introduction to Intelligent Control Systems
Step 2 or even Step 1 to come up with a simpler controller without violating
the performance requirements.

Another important aspect of implementation is the final adjustment, or

as often called the tuning, of the controller to improve performance by
compensating for the plant model uncertainties that are not accounted for
during the design process. Tuning is often done by trial and error and depends
very much on the experience and intuition of the control engineer.

1.1.4 Conventional Control System Approaches Problems

Finally, it should be stated that despite the success of conventional

control system approaches in dealing with many problems, they suffer from
many problems. These problems can be discussed below:
► Basically, these conventional approaches to control system design offer a
variety of ways to utilize information from mathematical models on how
to do good control. Because the plant model is always an approximation of
the plant, the effect of any difference between the plant and the model that
called plant model uncertainties on the performance of the controller will
not be known until the controller is applied to the plant. However, if these
uncertainties represent most of the un-modelled plant phenomena, its
representation in terms of mathematical equations is not possible.

► The conventional approaches do not take into account certain heuristic

information early in the design process, but sometimes use heuristics when
the controller is implemented (i.e., tuning is invariably needed since the
model used for the controller development is not perfectly accurate).
Chapter One
Introduction to Intelligent Control Systems
► Unfortunately, when using some approaches to conventional control, some
engineers become somewhat removed from the control problem (e.g.,
when they do not fully understand the plant and just take the mathematical
model as given), and sometimes this leads to the development of unrealistic
control laws.

► Sometimes in conventional control, useful heuristics are ignored because

they do not fit into the proper mathematical framework, and this can cause
problems.

These problems have given rise to the development of intelligent control

systems.

1.2 Concepts of Intelligent Control System

1.2.1 Definitions of Intelligent Control System

Intelligent Control (IC) was originally described by Fu in 1971, to

increase the flexibility and extend the range of then current automatic control
techniques. Since the early work by Fu, several application studies on
intelligent control have been developed. Although the term intelligent control
has been used in a variety of ways with different meanings in many
applications, most intelligent control systems have thus far attempted to
enhance conventional control methodologies to solve new challenging control
problems. To attain demanding goals for controlling complex systems, an
intelligent controller must cope with significant uncertainty that fixed-
parameter feedback controllers and adaptive controllers cannot handle.
Chapter One
Introduction to Intelligent Control Systems
Since Fu (1971) described intelligent control systems research as the
activities in the interaction of artificial intelligence and automatic control
systems, much effort has been made to formalize a unified definition of
intelligent control and classify it into a new discipline. Even though further
discussions and debates might be necessary for formalizing a unified
definition, the definition of intelligent control is likely to be found in the
following statements:

- "An intelligent control system has the ability to comprehend, reason and
learn about processes, disturbances and operating conditions" [Åström and
McAvoy,1992].
- "Intelligent control systems employ techniques that can sense and reason
about their environment and execute controls or commands in a flexible,
adaptive and robust manner" [Antsaklis and Passino,1993].

A task force was formed in May 1993 at the invitation of the Technical
Committee on Intelligent Control of the IEEE Control Systems Society to
characterize intelligent control systems, to distinguish them from
conventional approaches and to clarify the role of control in intelligent
systems. The final report extracted and summarized from the materials
contributed by the task force members on intelligent control states that:

- "An intelligent (control) system must be highly adaptable to significant

unanticipated changes, and so learning is essential. It must exhibit a high
degree of autonomy in dealing with changes. It must be able to deal with
significant complexity [IEEE, 1994].
Chapter One
Introduction to Intelligent Control Systems
1.2.2 Features of intelligent control System

Basically, the aim of intelligent control is to achieve and maintain high-

performance control for a system whose dynamics is:

1. Complex with ill-defined,

2. Nonlinear,
3. Time-varying,
4. Stochastic characteristics.
Although it is built on the background of classical control theory, intelligent
control provides a new discipline beyond conventional approaches. Until
now, intelligent control systems developed have had some unique features that
distinguish them from existing control systems. Those can be summarized as
follows:
1) Emulation of human experts with knowledge and inference engine,
2) Capability to learn from past experiences,

3) Identification of system changes.

1.2.3 Areas of intelligent control System

As already mentioned in the introduction section, fuzzy logic controller

(FLC), neural networks (NNS), fuzzy neural networks (FNNs), and genetic
algorithms (GAs) are important areas related to intelligent control. Whilst
techniques from these areas provide several key approaches to intelligent
control for complex systems, they are often interconnected to operating within
a hybrid structure. Therefore, the area of intelligent control is interdisciplinary
Chapter One
Introduction to Intelligent Control Systems
and the use of theories and methods from these four areas is helpful in building
an intelligent control system.

Because of their important roles in the development of intelligent control

systems, both the fuzzy logic controller and the neural networks are the
subjects of study and attention in this course. Thus, they will be described in
the remaining chapters.
.
15