2019 - Feedback Control Systems - Matlab - Simulink Approach - Farzin Asadi, Robert E. Bolanos, Jorge Rodriguez

Feedback Control Systems
The MATLAB®/Simulink® Approach

Synthesis Lectures on Control
and Mechatronics
Editors
Chaouki Abdallah, Georgia Institute of Technology
Mark Spong, University of Texas at Dallas
Feedback Control Systems: The MATLAB® /Simulink® Approach

Farzin Asadi, Robert E. Bolanos, and Jorge Rodríguez
2019
Mobile Robotics for Multidisciplinary Study

Carlotta A. Berry
2012
Control System Synthesis: A Factorization Approach, Part I

Mathukumalli Vidyasagar
2011
Control System Synthesis: A Factorization Approach, Part II

Mathukumalli Vidyasagar
2011
The Reaction Wheel Pendulum

Daniel J. Block, Karl J. Åström, and Mark W. Spong
2007
Copyright © 2019 by Morgan & Claypool
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in
any form or by any means—electronic, mechanical, photocopy, recording, or any other except for brief quotations
in printed reviews, without the prior permission of the publisher.
Feedback Control Systems: The MATLAB® /Simulink® Approach

www.morganclaypool.com
ISBN: 9781681735399 paperback

ISBN: 9781681735405 ebook
ISBN: 9781681735412 hardcover
DOI 10.2200/S00909ED1V01Y201903CRM005
A Publication in the Morgan & Claypool Publishers series

SYNTHESIS LECTURES ON CONTROL AND MECHATRONICS
Lecture #5
Series Editors: Chaouki Abdallah, Georgia Institute of Technology
Mark Spong, University of Texas at Dallas
Series ISSN
Print 1939-0564 Electronic 1939-0572
Feedback Control Systems
The MATLAB®/Simulink® Approach
Farzin Asadi
Kocaeli University, Kocaeli, Turkey
Robert E. Bolanos
Southwest Research Institute, Texas, U.S.A.
Jorge Rodríguez
Power Smart Control, Leganés, Spain
SYNTHESIS LECTURES ON CONTROL AND MECHATRONICS #5
M
&C Morgan & cLaypool publishers
ABSTRACT
Feedback control systems is an important course in aerospace engineering, chemical engineering,
electrical engineering, mechanical engineering, and mechatronics engineering, to name just a
few. Feedback control systems improve the system’s behavior so the desired response can be
acheived.
The first course on control engineering deals with Continuous Time (CT) Linear Time
Invariant (LTI) systems. Plenty of good textbooks on the subject are available on the market, so
there is no need to add one more. This book does not focus on the control engineering theories
as it is assumed that the reader is familiar with them, i.e., took/takes a course on control en-
gineering, and now wants to learn the applications of MATLAB® in control engineering. The
focus of this book is control engineering applications of MATLAB® for a first course on control
engineering.
KEYWORDS
closed-loop control, control, controller, feeback, PID controller, MATLAB® ,
Simulink®
vii
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
1 Introduction to MATLAB® . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Different Parts of MATLAB® . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 MATLAB’s Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 MATLAB’s Help System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 MATLAB’s Keywords . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.6 MATLAB’s Toolboxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.7 Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.8 Basic Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.9 Logical Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.10 Trigonometrics Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.11 Logarithmic and Exponential Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.12 Complex Numbers Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.13 Rounding Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.14 Remainder Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.15 Matrixes in MATLAB® . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.16 Commands: rand, ones, zeros, eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.17 The Colon Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.18 logspace and linspace Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.19 Operation with Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.20 Finding Polynomial Roots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.21 Product of Two Polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.22 Solution of Linear Systems of Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.23 Solution of Nonlinear Systems of Equations . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.24 Solving Differential Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.25 Differention and Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
viii
1.26 Drawing Diagrams with MATLAB® . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.27 Drawing the Graph of Collected Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
1.28 Drawing Two or More Data Simultanously . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
1.29 Drawing the Frequency Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2 Commonly Used Commands in Analysis of Control Systems . . . . . . . . . . . . . . 55

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.2 Defining the Transfer Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.3 Impulse Response of a System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.4 Step Response of a System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
2.5 Drawing the Frequency Response of a System . . . . . . . . . . . . . . . . . . . . . . . . . 78
2.6 Drawing the Nyquist Diagram of a System . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
2.7 Root Locus Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
2.8 Connecting the Systems Together . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3 Introduction to Simulink® . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.2 Running the Simulink® . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.3 Searching for Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.4 Example 1: Simulink® of a RLC Circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.5 Exporting the Results to Workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3.6 Example 2: Simulation in Transfer Functions . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.7 Example 3: Simulation of Electric Circuits with the Aid of Simscap Library 109
3.8 Example 4: Simulation of Nonlinear Systems . . . . . . . . . . . . . . . . . . . . . . . . . 123
3.9 Subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
3.10 Function Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
3.11 Example 5: Use of Function Block in the Simulink® Simulations . . . . . . . . . 128
3.12 Simulation of Discrete Time Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
3.13 Example 6: Simulation of Discrete Time Systems . . . . . . . . . . . . . . . . . . . . . 139
3.14 Example 7: Simulation of Discrete Time Systems . . . . . . . . . . . . . . . . . . . . . 144
3.15 Showing Two or More Signals Simultaneously . . . . . . . . . . . . . . . . . . . . . . . 145
3.16 Simulation of a Closed-Loop Control System in Simulink® . . . . . . . . . . . . . 151
4 Controller Design in MATLAB® . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
4.2 PID Controller Design in MATLAB® . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
ix
4.3 Tuning the PID Controllers Inside The Simulink® Environment . . . . . . . . . 172
4.4 Design of Linear Controllers with Sisotool (Control System Designer) . . . . 176
4.5 Loop Shaping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
4.6 Manual Controller Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
5 Introduction to System Identification Toolbox™ . . . . . . . . . . . . . . . . . . . . . . . 193

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
5.2 Illustrative Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Authors’ Biographies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

xi
Preface
This book is composed of five chapters. Here is a brief summary of these chapters:
Chapter 1: Introduction to MATLAB® . This chapter introduces the MATLAB® for
novices. The reader learns the fundamental concepts of MATLAB® .
Chapter 2: Commonly Used Commands in Analysis of Control Systems. This chapter
studies some of the most important commands of Control Systems Toolbox™ . Commands such
tf, zpk, ss, bode, nyquist, rlocus, impulse, step, feedback, seriesi, parallel, etc. are
studied in this chapter.
Chapter 3: Introduction to Simulink® . Simulink® is one of the most important tools
for simulation of dynamical systems. Simulink® is a graphical environment. The user makes the
model of system using plenty of ready to use blocks. So, thanks to the powerful library of blocks
Simulink® has, the user can simulate even complex systems easily without dealing with coding.
This chapter introduces the Simulink® with some examples.
Chapter 4: Controller Design in MATLAB® . Computers are an important part of mod-
ern engineering sciences. Nearly all the engineering designs are done with the aid of computers.
Control engineering is not an exception at all. MATLAB’s Control System Toolbox provides
plenty of functions to the design of control systems.
About 90% of controllers in industry are PID (Proportional-Integral-Derivative). Keep-
ing this in mind, the chapter starts with pidTuner command, a powerful command to design
PID controllers. The chapter finishes with the Control System Designer app.
Chapter 5: Introduction to System Identification Toolbox™ . System Identification
makes mathematical models for dynamical systems based on their input/output signals. An il-
lustrative example is studied in this chapter.

April 2019
xiii
Acknowledgments
The authors gratefully acknowledge MathWorks® support for this project.

April 2019
1
CHAPTER 1
Introduction to MATLAB®
1.1 INTRODUCTION
MATLAB® (MATrix LABoratory) is a multi-paradigm numerical computing environment and
proprietary programming language developed by MathWorks. MATLAB® allows matrix ma-
nipulations, plotting of functions and data, implementation of algorithms, creation of user in-
terfaces, and interfacing with programs written in other languages, including C, C++, C#, Java,
Fortran, and Python.
As of 2018, MATLAB® has more than 3 million users worldwide. MATLAB® users have
varied backgrounds in engineering, science, and economics. This chapter is a general introduc-
tion to MATLAB® .
Figure 1.1: Different parts of MATLAB® .

2 1. INTRODUCTION TO MATLAB®
1.2 DIFFERENT PARTS OF MATLAB®

The MATLAB® environment is shown in Fig. 1.1.
The main window of MATLAB® is composed of four parts.
Command Window
Command window is shown with number 1 in Fig. 1.1. The MATLAB® commands are written
in this window. After the command is written, the Enter key of the keyboard must be pressed
in order to run it (see Fig. 1.2).
Figure 1.2: Command window.
Work Space
Workspace window is shown with number 2 in Fig. 1.1. Variables defined inside the MATLAB®
environment are listed in Fig. 1.3.
Figure 1.3: Workspace.
Current Folder
Current folder window is shown with number 3 in Fig. 1.1. It shows the current working path.
You can set your desired path by clicking the “Browse for folder” icon (see Fig. 1.4).
Menu Bar
Menu bar is shown with number 4 in Fig. 1.1. It provides easy access to some of the widely
used operations (see Fig. 1.5). For instance, in order to enter the Simulink environment, the
user must type the word “simulink” in the command window. The user can get rid of writing the
word simulink by clicking the Simulink icon (Fig. 1.6) in the menu bar.
1.3. MATLAB’S EDITOR 3
Figure 1.4: Current folder.
Figure 1.5: Menu bar.
Figure 1.6: Simulink icon.
You can obtain the default layout, i.e., layout shown in Fig. 1.7, by clicking the “Default.”
1.3 MATLAB’S EDITOR

MATLAB® has a built-in editor which can be used for writing MATLAB® codes. In order to
run the editor, write the following command inside the command window:
>>edit
The editor is shown in Fig. 1.8.

You can save what you wrote inside the editor by using the “Save” icon (Fig. 1.9).
Figure 1.7: Layout icon.

1.3. MATLAB’S EDITOR 5
Figure 1.8: MATLAB’s editor.
Figure 1.9: Save the content of editor with the aid of save button.
Press the F5 key on the keyboard or click the Run icon in order to run the commands
written inside the editor (Fig. 1.10).
Figure 1.10: Run icon.
1.4 MATLAB’S HELP SYSTEM

You can press the F1 key on your keyboard in order to call the help window shown in Fig. 1.11.
Write the desired keywords in the “Search Documentation” box and press Enter key on
your keyboard.
MATLAB® has a command named help. For instance, you can see the help for dir
command by typing help dir inside the command window (Fig. 1.12).
You can use https://www.mathworks.com/help/ to get help as well. Write your key-
words inside the “Search Rxxxxx Documentation” box (Fig. 1.13).
1.5 MATLAB’S KEYWORDS

MATLAB® has 17 keywords, as shown in Table 1.1.
Keywords are shown in blue either in command window or MATLAB® editor (Fig. 1.14).
1.5. MATLAB’S KEYWORDS 7
Figure 1.11: Help window.
Figure 1.12: You can use the help command to obtain information about a command.
Figure 1.13: MATLAB’s online help.
Figure 1.14: Keywords are shown in blue.

1.6. MATLAB’S TOOLBOXES 9
Table 1.1: MATLAB® keywords
break case catch continue

else elseif end for
function global if otherwise
persistent return switch try
while
1.6 MATLAB’S TOOLBOXES

Toolboxes are a set of functions designed for a related purposes. They are sold as a package. Some
of the available toolboxes are shown in Table 1.2.
1.7 VARIABLES
Variable names are composed of English characters, numbers, or underline. The variable name
must start with an English character. For instance, “_myAge” or “1a” are NOT acceptable.
You can use the isvarname command in order to see if the selected name is acceptable or
NOT. For instance, if you write isvarname kp and press the Enter key, the MATLAB® shows
1. This means that the selected name (in this example “kp”) is acceptable. When MATLAB®
shows 0, it means that the name is NOT acceptable.
MATLAB® is a case-sensitive program, i.e., it is sensitive to small and capital letters. For
instance, “my_variable” and “My_variable” is NOT the same.
Some of the MATLAB’s simple commands are shown in Table 1.3.
MATLAB® does NOT show the command result on the screen if you put a semicolon at
the end of commands.
>>Dataspeed=115200;
>>Dataspeed=115200
Dataspeed =
115200
In order to see the variables defined variables, use the who command, i.e., write who in the
command window and press Enter key. If you want to see the size and number of bytes used for
variables, use the command whos.
Table 1.2: Some of MATLAB’s toolboxes
MATLAB Fixed-Point Designer Paralle Computing Toolbox

Simulink Fuzzy Logic Toolbox Phased Array Toolbox
Aerospace Blockset Global Optimization Toolbox Polyphase Bug Finder
Aerospace Toolbox HDL Coder Polyphase Code Prover
Anntenna Toolbox HDL Verifier Polyphase products for Ada
Audio System Toolbox IEC Certified Kit Power Train Blockset
Automated Driving System Toolbox Image Processing Toolbox RF Blockset
Bioinformatics Toolbox Instalation, Licensing and Activation RF Toolbox
Communications System Toolbox Instrument Control Toolbox Risk Management Toolbox
Computer Vision Toolbox LTE System Toolbox Robotic Toolbox
Control System Toolbox Mapping Toolbox Robust Control Toolbox
Curve Fitting Toolbox MATLAB Coder Signal Processing Toolbox
Data Acquisition Toolbox MATLAB Compiler SimBiology
Database Toolbox MATLAB Compiler SDK SimEvents
Datafeed Toolbox MATLAB Distributed Computing Simscape
Toolbox
DO Qualification Kit (for DO-178) MATLAB Production Server Simscape Driveline
DSP System Toolbox MATLAB Report Generator Simscape Electronics
Econometrics Toolbox Predictive Control Toolbox Simscape Fluids
Embeded Coder Model Base calibration Toolbox Simscape Multibody
Filter Design HDL Coder Nueral Network Toolbox Simscape Multibody Link
Financial Instruments Toolbox OPC Toolbox Simscape Power Systems
Financial Toolbox Optimization Toolbox Simulink 3D Animation
Simulink 3D Animation Simulink PLC Coder Symbolic Math Toolbox
Simulink Code Inspector Simulink Real Time System Identification Toolbox
Simulink Coder Simulink Report Generator Trading Toolbox
Simulink Control Design Simulink Test Vehicle Network Toolbox
Simulink Design Optimization Simulink Verification and Validation Wavelet Toolbox
Simulink Design Verifier Spreadsheet link WLAN System Toolbox
Simulink Desktop Real-Time Stateflow
1.8. BASIC OPERATORS 11
Table 1.3: Some of MATLAB’s simple commands
Commands Explanation
>>x=2018 This command asked MATLAB to make a variable named “x”. x is 1×1
matrix and contains value 2018.
>>x This command asks MATLAB to show the content of variable “x”.
>>x=x+1 This command asks MATLAB to increase the value of variable x by 1.
>>Name= ’John’ This command asks MATLAB to make a variable named “Name”.
“Name” is 1×4 matrix contains “John” word.
You can clear the screen by using the command clc. In order to clear all the variables, use
the command clear all. If you want to clear a specific variable use the command clear. You
must write the variable name after the clear.
1.8 BASIC OPERATORS

MATLAB’s basic operators are shown in Table 1.4.
Table 1.4: MATLAB’s basic operators
Operator Explanation Operator Explanation

+ Addition ^ Matrix power
- Subtraction .* Elementwise multiplication
* Matrix Multiplication ./ Elementwise right division
/ Division .\ Elementwise left division
. Decimal point .^ Elementwise power
\ Matrix left division .’ Transpose
= Assignment
Example 1.1
>>A=[1 2 3];
>>B=[4 5 6];
>>A./B
ans =
0.2500 0.4000 0.5000
>>A.^B
ans =
1 32 729
>>A.*B
ans =
4 10 18
1.9 LOGICAL OPERATORS

MATLAB’s logical operators are shown in Table 1.5.
1.10 TRIGONOMETRICS FUNCTIONS

MATLAB’s trigonometric functions are shown in Table 1.6.
Example 1.2
Calculate sin.x/4 3 cos.x/ for x D 30ı .
Solution:
First way:
>>x=30*pi/180;
>>sin(x)^4-3*cos(x)
ans=
-2.5356
1.10. TRIGONOMETRICS FUNCTIONS 13
Table 1.5: MATLAB’s logical operators
Operator Explanation Example

== or eq When left and right operand are >> 5==6
equal, it returns 1. Otherwise it ans =
returns 0. 0
>> 5==5
ans
1
~= or ne When left and right operand >> 5~=6
are NOT equal, it returns 1. ans =
Otherwise it returns 0. 1
>> 5~=5
ans =
0
> or gt (greater than) When the inequality is satisfied, it >> 5>=6
>= or ge (greater than or equal) returns 1. Otherwise it returns 0. ans =
< or lt (less than) 0
<= or le (less than or equal) >> 5<=15
ans =
1
& or and When the two operand are true, it >>and(eq(3,3),5>2)
returns 1. Otherwise it returns 0. and =
1
| or or When at least one of the two >>eq(5,5)|3>12
operand are true, it returns 1. ans =
Otherwise it returns 0. 1
~ or NOT Returns the complement of the >>~eq(5,5)
logical expression. ans =
0
>>NOT(eq(5,5))
ans =
0
Table 1.6: MATLAB’s trigonometric functions (Continues)
Function Explanation Example

sin Calculates the sine of an angle. Operand must be >>sin(1)
given in Radyans. ans =
0.8415
sind Calculates the sine of an angle. Operand must be >>sind(30)
given in Degrees. ans =
0.5000
asin Calculates the arc sine. Result is in Radyans. >>asin(0.5)
ans =
0.5236
asind Calculates the arc sine. Result is in Degrees. >>asind(.5)
ans =
30.0000
cos Calculates the cosine of an angle. Operand must >>cos(1)
be given in Radyans. ans =
0.5403
cosd Calculates the cosine of an angle. Operand must >>cosd(.5)
be given in Degrees. ans =
1.0000
acos Calculates the arc cosine. Result is in Radyans. >>acos(.5)
ans =
1.0472
acosd Calculates the arc cosine. Result is in Degrees. >>acosd(.5)
ans =
60.0000
tan Calculates the tangent of an angle. Operand must >>tan(1)
1.5574
tand Calculates the tangent of an angle. Operand must >>tand(60)
1.7321
1.10. TRIGONOMETRICS FUNCTIONS 15
Table 1.6: (Continued) MATLAB’s trigonometric functions (Continues)
atan Calculates the arc tangent. Result is in Radyans. >>atan(1)

ans =
0.7854
atand Calculates the arc tangent. Result is in Degrees. >>atand(30)
ans =
88.0908
sec Calculates the secant of an angle. Operand must >>sec(1)
1.8508
secd Calculates the secant of an angle. Operand must >>secd(10)
1.0154
asec Calculates the arc secant. Result is in Radyans. >>asec(30)
ans =
1.5375
asecd Calculates the arc secant. Result is in Degrees. >>asecd(30)
ans =
88.0898
csc Calculates the cosecant of an angle. Operand must >>csc(1)
1.1884
cscd Calculates the cosecant of an angle. Operand must >>cscd(1)
57.2987
acsc Calculates the arc cosecant. Result is in Radyans. >>acsc(1)
ans =
1.5708
acscd Calculates the arc cosecant. Result is in Degrees. >>acscd(1)
ans =
90
cot Calculates the cotangent of an angle. Operand >>cot(1)
must be given in Radyans. ans =
0.6421
Table 1.6: (Continued) MATLAB’s trigonometric functions
cotd Calculates the cotangent of an angle. Operand >>cotd(1)

must be given in Degrees. ans =
57.2900
acot Calculates the arc cotangent. Result is in Radyans. >>acot(1)
ans =
0.7854
acotd Calculates the arc cotangent. Result is in Degrees. >>acotd(60)
ans =
0.9548
Second way:
>>x=30;
>>sind(x)^4-3*cosd(x)
ans=
-2.5356
1.11 LOGARITHMIC AND EXPONENTIAL FUNCTIONS

MATLAB’s logarithmic and exponential functions are shown in Table 1.7.
NOTE: In MATLAB® , 1.234e5 equals 1:234 105 . See Table 1.8 for more examples.
1.12 COMPLEX NUMBERS FUNCTIONS

MATLAB’s functions related to complex numbers are shown in Table 1.9.
1.13 ROUNDING FUNCTIONS

MATLAB’s rounding functions are shown in Table 1.10.
1.14 REMAINDER FUNCTIONS

MATLAB’s remainder functions are shown in Table 1.11.
1.14. REMAINDER FUNCTIONS 17
Table 1.7: MATLAB’s logarithmic and exponential functions

exp Calculates . >>exp(2)
(Note: = 2.7182) ans =
7.3891
log Calculates the log ( >>log(2.7182)
ans =
1.0000
log10 Calculates the log10 ( >>log10(100)
ans =
2
log2 Calculates the log2 ( >>log2(4)
ans =
2
sqrt Calculates the . >>sqrt(4)
ans =
2
power Calculates , i.e., power (a,b) = . >>power(2,3)
ans =
8
Table 1.8: Conversion of simple expression to MATLAB®
Mathematics MATLAB
1234 >>12^34
1 -3 >>(1/2)^(-3)
2
123 × 10-4 >>123*10^-4
or
>>123e-4
3
124 >>12^(3/4)
Table 1.9: MATLAB’s functions related to complex numbers

complex This command is used to produce a complex >>complex(2,3)
number, i.e., = + × where ans =
= √–1. 2.0000 + 3.0000i
angle Returns the angle of complex number. It returns >>angle(2+4i)
the angle in Radyans. ans =
1.1071
conj Returns the complex conjugate of a complex >>conj(2+5i)
number. ans =
2.0000 – 5.0000i
imag Returns the imaginary part of a complex number. >>imag(4+9i)
ans =
9
real Returns the real part of a complex number. >>real(4+9i)
ans =
4
isreal When the input argument is a real number, i.e., >>isreal(2+3i)
imaginary part is zero, it returns 1. Otherwise it ans =
returns 0. 0
>>isreal(4)
ans =
1
abs Calculates the absolut value. When the input ar- >>abs(3-4*i)
gument is a complex number it returns the mag- ans =
nitude of the number, i.e., 5
+ × = √ 2 + 2 ).
1.14. REMAINDER FUNCTIONS 19
Table 1.10: MATLAB’s rounding functions

fix Round toward zero. >>fix(4.6)
ans =
4
floor Round toward negative infinity. >>floor(8.4687)
ans =
8
ceil Round toward positive infinity. >>ceil(4.4)
ans =
5
round Round to nearest decimal or integer. >>round(4.55)
ans =
5
Table 1.11: MATLAB’s remainder functions

mod Remainder after division (modulo operation). >>mod(17,3)
ans =
2
rem Remainder after division. >>rem(17,3)
ans =
2
sign Sign function (signum function). >>sign(4)
ans =
1
>>sign(-4)
ans =
-1
1.15 MATRIXES IN MATLAB®

MATLAB® considers everything as matrices. For instance, A, B, C, D are 11 matrices and E
is a 114 matrix.
>>A=123;
>>B=-456.7;
>>C=12e-3;
>>D=12+34*i;
>>E='press any key...';
Table 1.12 shows how to define larger matrices.
Table 1.12: Entering a matrix in MATLAB® . (Note: in order to read the array in the 3rd row
and 2nd column of matrix F, you write, F(3,2).)
Mathematics MATLAB
1 2 3 >>F=[1 2 3;4 5 6;7 8 9]
or
>>F=[1,2,3;4,5,6;7,8,9]
1.16 COMMANDS: RAND, ONES, ZEROS, EYE

• rand: produce a matrix with uniformly distributed random arrays. For instance,
A=rand(1,3) makes a 1 3 matrix with random arrays. Each array is between 0 and 1.
• ones: reate array of all ones. For instance, ones(k), creates a k k matrix which all the
arrays are 1.
• zeros: create array of all zeros. For instance, zeros(k), creats a k k matrix which all
the arrays are 0.
• eye: makes identity matrix. For instance, eye(4) create a 4 4 identity matrix.
1.17. THE COLON OPERATOR 21
1.17 THE COLON OPERATOR
You can use the colon operator (:) to creat matrices. Here is the syntax:
f D Initial value:Change steps:Final value
Example 1.3
>>n=[1:2:10]
n =
1 3 5 7 9
>>n(2:4)
ans =
3 5 7
1.18 LOGSPACE AND LINSPACE COMMANDS

The logspace function generates logarithmically spaced vectors. For instance,
y=logspace(a,b) generates a row vector y of 50 logarithmically spaced points between
decades 10a and 10b . y=logspace(a,b,n) generates n points between decades 10a and 10b .
logspace command is especially useful for creating frequency vectors.
The linspace function generates linearly spaced vectors. For instance, y=linspace(a,b)
generates a row vector y of 50 linearly spaced points between decades a and b .
y=linspace(a,b,n) generates n points between decades a and b .
1.19 OPERATION WITH MATRICES

In order to obtain the size of a matrix, use the size command. For instance, size(A) returns
the number of rows and number of colums of a matrix named A.
diag(A) returns the arrays on the main diagonal of square matrix A.
det(A) returns the determinant of square matrix A.
inv(A) or A^-1 returns the inverse of square matrix A.
eig(A) returns the eigen values of square matrix A. [v,d]=eig(A) returns the eigen vectors as
well.
rank(A) returns the rank of matrix A.
transpose(A) returns the transpose of matrix A.
ctranspose(A) or A' returns the transpose of matrix A. The complex conjugate transpose of a
matrix interchanges the row and column index for each element, reflecting the elements across
the main diagonal. The operation also negates the imaginary part of any complex numbers. For
example, if B = A' and A(1,2) is 1+1i, then the element B(2,1) is 1-1i.
1.20 FINDING POLYNOMIAL ROOTS

Consider a polynomial function as follows:
f D x 4 C 2x 3 C 3x 2 C 4x C 5:
Enter the polynomial function coefficients as a vector f.
>>f=[1 2 3 4 5]
The roots can be find with the following command:
>>roots(f)
You can use the command polyval in order to obtain the value of a polynomial function at a
specific point. For instance, in order to calculate the polynomial f at x=0 you write:
>>polyval(f,0)
1.21 PRODUCT OF TWO POLYNOMIALS

You can calculate the product of two polynomials with the aid of conv command. For instance,
to find the product of f D x 4 C 2x 3 C 3x 2 C4x C 5 and g D x 3 C 5, we write:
>>f=[1 2 3 4 5];
>>g=[1 0 0 5];
>>conv(f,g)
1.22. SOLUTION OF LINEAR SYSTEMS OF EQUATIONS 23
ans =
1 2 3 9 15 15 20 25
So the product of afforementioned polynomials is, x 7 C 2x 6 C 3x 5 C 9x 4 C 15x 3 C 15x 2 C

20x C 25.
1.22 SOLUTION OF LINEAR SYSTEMS OF EQUATIONS

You can use the linear system of equations with the aid of MATLAB® . See Table 1.13.
Table 1.13: Solving a linear system of equations in MATLAB® . (Note: instead of BnA you can
use inv(A)*B or A O -1*B.)
Mathematics MATLAB
>>A=[1 2 3;4 5 6;7 8 8];
>>B=[1;2;3];
1 2 3 1
>>B\A
4 5 6 × = 2
ans=
7 8 9 3
-0.3333
0.6667
0
1.23 SOLUTION OF NONLINEAR SYSTEMS OF

EQUATIONS
MATLAB® can solve the nonlinear system of equations as well. fzero function can be used for
this purpose. The syntax is:
fzero.‘Function Name’,initial value/
You must prepare a MATLAB® function that contains the desired equation. The structure of a
MATLAB® function is shown bellow:
function output_argument D FunctionName.InputArgument/

output_argument D DesiredEquation
end
study the following example to see how to use the fzero command.
Example 1.4
Find the root of f D x 4 C 3x 3 C 3x 2 C 14x C 5 with the aid of fzero command.
Solution:
In order to prepare the MATLAB® function containing the non linear equations, enter the
editor environment (Fig. 1.15).
Figure 1.15: You can enter the editor by typing edit in the command line.
Write the following codes inside the editor (Fig. 1.16).
Figure 1.16: The function contains the equations.
function y=f(x)
y=x^4+3*x^3+3*x^2+14*x+5;
end
1.24. SOLVING DIFFERENTIAL EQUATIONS 25
Save the file as f.m. When you save a function, you must use the same name as the function
name. Since we call the MATLAB® function f, we must save it as f.m. Type the following
command in command prompt of MATLAB® :
>>fzero(`f',0)
ans=
-0.3776
NOTE: Set the “Current Folder” window of MATLAB® to where you saved f.m, other-
wise MATLAB® cannot find the f.m.
1.24 SOLVING DIFFERENTIAL EQUATIONS

You can solve Ordinary Differential Equations (ODE) and even Partial Differential Equations
(PDE) with the aid of MATLAB® .
See the following examples to learn different ways of solving ODEs.
Example 1.5
Solve dy.t
dt
/
C 4y .t/ D e t ; y .0/ D 1 with MATLAB® .
Solution:
See Fig. 1.17.
dy.t/
Figure 1.17: Solving the dt
C 4y .t/ D e t ; y .0/ D 1. (Note: syms defines symbolic vari-
ables.)
Example 1.6
2
Solve 2 d y.t/
2 C dy.t /
dt
C 11y .t/ D e t ; y .0/ D 1; y 0 .0/ D 0 with MATLAB® .
dt
Solution:
See Fig. 1.18.
2 y.t/
Figure 1.18: Solving the 2 d C dy.t/
C 11y .t/ D e t ; y .0/ D 1; y 0 .0/ D 0.
dt 2 dt
dn
This example used the “Dn” operator instead of “diff” command(Dn(.)D dt n
.:/).
Example 1.7
Solve dy.t/
dt
D ty with MATLAB® .
Solution:
See Fig. 1.19.
dy.t /
Figure 1.19: Solving the dt
D ty .
1.24. SOLVING DIFFERENTIAL EQUATIONS 27
Example 1.8
Solve . dy.t
dt
/
C y.t //2 D 1; y .0/ D 0 with MATLAB® .
Solution:
See Fig. 1.20.
Figure 1.20: Solving the . dy.t

dt
/
C y.t //2 D 1; y .0/ D 0.
Example 1.9
d 2 y.x/
Solve D cos .2x/ y; y .0/ D 1; y 0 .0/ D 0 with MATLAB® .
dx 2
Solution:
See Fig. 1.21.
d 2 y.x/
Figure 1.21: Solving the D cos .2x/ y; y .0/ D 1; y 0 .0/ D 0.
dx 2
1.25 DIFFERENTION AND INTEGRATION

Examples for differention and integration in MATLAB® are shown in Table 1.14.
Table 1.14: Some examples for differention and integration in MATLAB®
Mathematics MATLAB
3– >>syms x
= 2+ 8) >>diff(‛(x^3-x)/(x^2+8)’)
=?
= 3+ 2+ >>syms x y
>>f=x^3+3*y^2+4*x*y
=?
>>diff(f,y)
3– >>syms x
2+ 8 >>int((x^3-x)/(x^2+8))
1 3 >>syms x
>>int((x^3)/(x^2+8),x,0,1)
1.26. DRAWING DIAGRAMS WITH MATLAB® 29
1.26 DRAWING DIAGRAMS WITH MATLAB®

You can use the plot(x,y) command to draw diagrams. See the following examples.
Example 1.10
>>x=[0:0.01:2*pi];
>>y=sin(x);
>>plot(x,y)
See Fig. 1.22.
Figure 1.22: Output of Example 1.10.

Example 1.11
>>x=[0:0.01:2*pi];
>>y1=sin(x);
>>y2=cos(x);
>>plot(x,y1,x,y2)
See Fig. 1.23.

You can select the desired line style, color or marker type for the drawn diagram. To do
this, you use the plot(x,y,'?') command. ? shows the desired line style/color/marker type.
Line style, color or marker type is set according to Tables 1.15, 1.16, and 1.17.
Table 1.15: Line style table
Command Line style

- Solid line
-- Dashed line
: Dotted line
-. Dotted and
dashed line
Table 1.16: Colors table
Command Color
r red
g green
b blue
c cyan
m magenta
y yellow
k black
w white
Table 1.17: Available markers
.
+
*
O
x
s
d
v
^
<
>
Table 1.18 shows the commands in relation to the plot command.
Table 1.18: Some of the commands related to plot command
Command Explanation
grid Display or hide axes grid lines
title(‘’) Add title.
xlabel(‘’) Label X axis.
ylabel(‘’) Label Y axis.
text(x,y,’’) Add text descriptions to data points
See the following examples.
Example 1.12
>>x=[0:0.01:2*pi];
>>y=sin(x+pi/4);
>>plot(x,y,'b--')
>>xlabel(`Time(s)')
>>ylabel(`Voltage Values(V)')
>>grid on
See Fig. 1.24.
You can draw the bargraph, stairstep graph, and pie chart with the aid of bar, stair, and
pie commands, respectively.
Example 1.13
>>a=[0 1 3 7 9];
>>b=[2 1 4 2 3];
>>figure(1)
>>bar(a,b)
>>stairs(a,b)
>>pie(a)
See Figs. 1.25, 1.26, and 1.27.



1.27. DRAWING THE GRAPH OF COLLECTED DATA 37
1.27 DRAWING THE GRAPH OF COLLECTED DATA
MATLAB® can be used to draw the graph of data collected in lab. Both linear and logarithmic
graphs can be draw with the aid of MATLAB® . The following example shows how to do this.
Example 1.14
Table 1.19 shows the current-voltage for a resistor. We want to draw the plot of this data.
Table 1.19: Measured voltage and current for resistor R1
V (Volt) I (Amper)
0.499 0.10
0.985 0.20
1.508 0.31
1.969 0.41
2.528 0.53
2.935 0.61
3.481 0.73
3.971 0.83
4.486 0.94
4.960 1.04
5.502 1.15
6.007 1.26
6.60 1.38
Enter the obtained data to MATLAB® . See Fig. 1.28.
Figure 1.28: Entering the measured data.
Use the plot command to draw the graph (Fig. 1.29). The result is shown in Fig. 1.30.
If you need smaller grid axes, use the command grid minor. Results are shown in
Figs. 1.31 and 1.32.
Figure 1.29: Drawing the data.
Figure 1.30: Graph of measured data.

Figure 1.31: Using smaller grid axes.
Figure 1.32: Smaller axes grids are used.

You can use the command shown in Fig. 1.33 in order to show the data points with a
marker. The result is shown in Fig. 1.34.
Figure 1.33: Showing the data points with red stars.
Figure 1.34: Graph of measured data. Data points are shown with red stars.
Use “xlabel” and “ylabel” to add descriptive texts to the axes (Figs. 1.35 and 1.36).
Figure 1.35: Adding labels to axes.
Figure 1.36: Labels are added to axes.

You can add text to axes with the aid of Insert menu (see Fig. 1.37) as well. After Insert
menu is opened, click on “X Label” or “Y Label” to add text to x axis and y axis, respectively.
Figure 1.37: You can add labels to the axes with the aid of insert menu.
You can save the drawn graph as a graphic file. To do this use File>Save As… See
Fig. 1.38.
Figure 1.38: You can export the drawn graph as a graphic file.
Select the desired graphic file format and save the file. See Fig. 1.39.
Figure 1.39: Selection of desired format.
1.28 DRAWING TWO OR MORE DATA SIMULTANOUSLY

Drawing a graph of two or more data simultanously is so common. Drawing such graphs in
MATLAB® is shown with an example.
Consider we have measured the data shown in Table 1.20 for another resistor.
We want to show the I–V relationship of the first resistor (Table 1.19) and the second
resistor (Table 1.20) on the same graph. Codes shown in Fig. 1.40 show how to do this. The
result is shown in Fig. 1.41.
Use Insert>Legend to show which plot belong to which resistor (Fig. 1.42).
After you clicked the “Legend,” MATLAB® adds the default legend to the plot (Fig. 1.43).
Double-click the data1, data2, data3, and data4 to change them into a more convinient
text. You can drag the legend box to desired place, as well (Fig. 1.44).
1.28. DRAWING TWO OR MORE DATA SIMULTANOUSLY 45
Table 1.20: Measured voltage and current for resistor R2
V (Volt) I (Amper)
0.579 0.10
0.978 0.17
1.598 0.28
1.976 0.34
2.496 0.43
2.953 0.51
3.458 0.60
4.068 0.71
4.450 0.78
4.917 0.86
5.35 0.93
5.75 1.01
6.37 1.11
6.60 1.15
Figure 1.40: You draw two or more plots on the same axes simultanously with the aid of hold
on command.
Figure 1.41: Result of code shown in Fig. 1.40.

Figure 1.42: Adding legend to the drawn graph.

Figure 1.43: Default legend is added to the graph.

Figure 1.44: Editing and replacing the default legend.

1.29 DRAWING THE FREQUENCY RESPONSE
Frequency response plots are drawn on the logarithmic scale. You can draw the logarithmic plots
with the aid of semilogx, semilogy, and loglog commands.
Drawing the frequency response plot is usually done with the aid of semilogx command.
Since only the x axis of frequency response is in logarithmic scale.
The procedure of drawing frequency response plots is shown with an example. Consider
a data as shown in Table 1.21. This is the frequency response of the simple RC circuit shown in
Fig. 1.45.
Table 1.21: Frequency response data for circuit shown in Fig. 1.45
Magnitude Phase (
∢
( ( )
Frequency (Hz)
( ) (Degrees)
1 1.000 -0.36
10 0.998 -3.60
20 0.992 -7.16
50 0.954 -17.44
100 0.847 -32.13
150 0.728 -43.30
200 0.623 -51.48
250 0.537 -57.51
300 0.469 -62.05
350 0.414 -65.54
400 0.370 -68.30
450 0.333 -70.51
500 0.303 -72.34
550 0.278 -73.85
600 0.256 -75.14
We enter the data to MATLAB® (Fig. 1.46). ˇ ˇ

ˇ ˇ
We want to draw the response in dB. To do this we must calculate the 20 log10 ˇ VVino .j!/
.j!/ ˇ
.
Commands shown in Fig. 1.47 draw the frequency response of data given in Table 1.21. The
We need to draw the phase diagram as well. We add the command shown in Fig. 1.49 to
draw the phase diagram. The result is shown in Fig. 1.50.
1.29. DRAWING THE FREQUENCY RESPONSE 51
1k
+
+
~
–
Vin 1uF Vout
–
Figure 1.45: Simple lowpass RC circuit.
Figure 1.46: Entering the magnitude and phase to MATLAB® .
Figure 1.47: Drawing the magnitude graph.
You can add a descriptive title to the graph. To do this, click the Insert>Title and type
the desired title (Fig. 1.51).
Figure 1.48: The magnitude graph.
Figure 1.49: Drawing the phase graph.

1.29. DRAWING THE FREQUENCY RESPONSE 53
Figure 1.50: The magnitude and phase graph.

Figure 1.51: Addition of a title to the drawn plot.

55
CHAPTER 2
Commonly Used Commands in

Analysis of Control Systems
2.1 INTRODUCTION
Chapter 1, was a general introduction to MATLAB. This chapter, introduces some of the im-
portant commands used in analysis of control systems.
2.2 DEFINING THE TRANSFER FUNCTION

The procedure is shown with two examples.
Example 2.1
We want to define the following transfer function in MATLAB.
100
H .s/ D
s2 C 6s C 100
>>H=tf([100],[1 6 100]);
As another way, you can write:
>> s=tf('s');
>> H=100/(s^2+6*s+100)
Example 2.2
To define following transfer function,
s 2 6s
I.s/ D
s 3 C 6s C 100
>>I=tf([100 -6 0],[1 0 6 100]);

56 2. COMMONLY USED COMMANDS IN ANALYSIS OF CONTROL SYSTEMS
Example 2.3
To define the following transfer function,
sC1
J.s/ D :
.s C 2/ .s C 3/ .s C 4/
One may do some simple manipulation and write,
sC1 sC1
J.s/ D D 3 2
;
.s C 2/ .s C 3/ .s C 4/ s C 9s C 26s C 24
and then use the following command:
>>J=tf([1 1],[1 9 26 24]);
However there is a simpler solution using zpk command:
>>J=zpk([-1],[-2 -3 -4],1);
The zpk command takes three inputs: the first input is a vector contain transfer function zeros,
the second is the vector contains transfer function poles, and the third one is a gain. Since J,
has only one zero places at 1, the zero vector contains only one array, 1. The J has poles
at 2, 3, and 4. So, second input of zpk command is [ 2 3 4]. The gain zero since
sC1 sC1
J .s/ D .sC2/.sC3/.sC4/ D 1 .sC2/.sC3/.sC4/ . To model K .s/ D 5 .sC2/.ss 13/.sC4/ one must
use the command >>K=zpk([+1],[-2 +3 -4],-5).
Figure 2.1 shows how to transfer a zpk model to tf model.
You can define state space models as well. Use ss command to define state space models.
See Fig. 2.2 to see an example.
You can obtain the transfer function associated with the state space model with the aid of
tf command (Fig. 2.3).
You can obtain the location of polse and zeros with the aid of pole and zero commands,
respectively (Figs. 2.4 and 2.5).
You can obtain a graphical plot of location of poles and zeros with the aid of pzmap com-
mand. pzmap shows the location of poles with a mark and location of zeors with an open circle.
For instance, you can draw the pole-zero map of L with the aid of the following command. The
>>pzmap(L)
A system is stable if all the poles lies in the open Left Half Plane (LHP), i.e., the left half
plane without the imaginary access. You can use the command isstable to learn whether or
not the system is stable.
2.2. DEFINING THE TRANSFER FUNCTION 57
Figure 2.1: Converting a zpk model to tf model.

Figure 2.2: Defining the state space model.

Figure 2.3: Converting a state space model to transfer function model.

Figure 2.4: Obtaining the poles and zeros. H, is a transfer function model.
Figure 2.5: Obtaining the poles and zeros. L, is a state space model.
Figure 2.6: Pole-zero map of L.
>>H=tf([100],[1 6 100]);
>>isstable(H)
ans=
1
2.3 IMPULSE RESPONSE OF A SYSTEM
Impluse response of a system can be drawn with the aid of impulse command.
Example 2.4
Draw the impulse response of following system. The result is shown in Fig. 2.7.
100
H .s/ D
s 2 C 6s C 100
>>H=tf([100],[1 6 100]);
>>impluse(H)
>>grid on
You can ask MATLAB to show peak response and settling time. To do this, right-click
on the plot. Click on the Characteristics and select your desired option. For instance, we select
the Peak Response (Fig. 2.8).
After you clicked the Peak Response, the MATLAB shows the Peak of graph with a point
(Fig. 2.9).
After you clicked on the shown point, MATLAB adds a label and shows the coordimnate
of the point with a label (Fig. 2.10).
You can use the Data Cursor icon shown in Fig. 2.11 to add cursor to the graph. First
click the Data Cursor icon and then click a point on the graph.
Using the cursor, you can read the coordinate of desired point easily (Fig. 2.12).
You can slide the cursor to desired point. You can slide the cursor more easily if you set
the Selection Style to Mouse Position (Fig. 2.13). To see the Selection Style, right-click on the
graph.
You can zoom in/out the drawn graph with the aid of a magnifier icon (Fig. 2.14).
You can export the drawn graph as a graphic file. To do this, click theSave as… (Fig. 2.15).
Type your desired name and Select the desired graphic file (i.e., bmp, jpg, png) and export
the graphic (Fig. 2.16).
2.3. IMPULSE RESPONSE OF A SYSTEM 63
Figure 2.7: Impulse response of H.

Figure 2.8: Obtaining the impulse response characteristics.

Impulse Response
7
3
Amplitude
-1
-2
-3
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Time (s)
Figure 2.9: Peak of the graph is shown with a filled blue circle.
Figure 2.10: Coordinate of the peak point.

Figure 2.11: Data cursor icon.

Figure 2.12: You can read the coordinate of desired point with the aid of added cursor. To remove
the shown box, left-click on it and press the delete key on the keyboard.
Figure 2.13: You can move the cursor more smoothly if you select the mouse position.
Figure 2.14: You can zoom in/out with the aid of a magnifier.
Figure 2.15: You can save the drawn plot with the aid of save as ….
Figure 2.16: Selection of desired output file format.
You can draw the impulse response in a specific interval if you want. For instance, to draw
the impulse response in the [.2,1.8] interval you type:
>>H=tf([100],[1 6 100]);
>>impluse(H,[.2:.01:1.8])
>>grid on
NOTE: .01, is the time step. The above code asks MATLAB to calculate the impulse
response at t D 0:2; 0:21; 0:22; 0:23; : : : ; 1:8 and draw them. The result is shown in Fig. 2.17.
2.4. STEP RESPONSE OF A SYSTEM 73
Impulse Response
6
Amplitude
2
-1
-2
-3
0.4 0.6 0.8 1 1.2 1.4 1.6
Time (s)
Figure 2.17: Drawing the impulse response for Œ0:2; 1:8.
2.4 STEP RESPONSE OF A SYSTEM

Step response of a system can be drawn with the aid of step command.
Example 2.5
Draw the step response of following system. The result is shown in Fig. 2.18.
100
H .s/ D
s 2 C 6s C 100
>>H=tf([100],[1 6 100]);
>>step(H)
>>grid on
Figure 2.18: Step response of H.
You can draw the step response in a specific interval if you want. For instance, to draw the
step response in the [.2,1.8] interval you type:
>>H=tf([100],[1 6 100]);
>>step(H,[.2:.01:1.8])
>>grid on
NOTE: .01, is the time step. The above code asks MATLAB to calculate the step response
at t D 0:2; 0:21; 0:22; 0:23; : : : ; 1:8 and draw them. The result is shown in Fig. 2.19.
Step Response
1.4
1.3
1.2
Amplitude
1.1
0.9
0.8
0.4 0.6 0.8 1 1.2 1.4 1.6
Time (s)
Figure 2.19: Drawing the impulse response for [0.2,1.8].
You can ask MATLAB to measure the important characteristics of the response and shows
them. To do this, right-click on the step response plot and click characteristics. Select desired
option from the shown menu. See Fig. 2.20.
For instance, we select the Rise Time from the shown menu. The result is shown in
Fig. 2.21. Rise time is the time it takes for the response to rise from 10–90% of the steady-
state response.
You can use the command stepinfo to see the step response characteristics (Fig. 2.22).
Figure 2.20: Obtaining the characteristics of the drawn graph.

Figure 2.21: The rise time of the system is 0.132.

Figure 2.22: Step response characteristics.
2.5 DRAWING THE FREQUENCY RESPONSE OF A

SYSTEM
You can draw the frequency response of a system with the aid of bode command (Fig. 2.23). For
example, the bode diagram of system of Example 2.1 can be drawn with the aid of following
command:
>>bode(H),grid on
If you want to draw the frequency response of a system in a specific interval, you must de-
fine the desired interval. You can use the bode(sys,w) where w is the vector of desired frequency
range.
h i if you want to draw the frequency response of system in Example 2.1 for
For instance,
Rad Rad
100 s ; 1000 s interval, you can use the following code. The result is shown in Fig. 2.24.
>>w=logspace(log10(100),log10(1000))
>>bode(H,w)
>>grid on
You can measure the important characteristics of the drawn graph by right-clicking on it
and select the Characteristics (Fig. 2.25).
2.5. DRAWING THE FREQUENCY RESPONSE OF A SYSTEM 79
Figure 2.23: Bode plot of H.

Bode Diagram
-30
-40
Magnitude (dB)
-50
-60
-70
-80
-176
-177
Phase (deg)
-178
-179
-180
10-2 103
Frequency (rad/s)
h i
Figure 2.24: Bode plot of H for 100 Rad
s
; 1000 Rad
s
interval.
2.6. DRAWING THE NYQUIST DIAGRAM OF A SYSTEM 81
Figure 2.25: Obtaining the characteristics of the plot.
2.6 DRAWING THE NYQUIST DIAGRAM OF A SYSTEM

You can draw the Nyquist diagram of a system with the aid of nyquist command. For example,
the Nyquist diagram of system of Example 2.1 can be drawn with the aid of following command.
The result is shown in Fig. 2.26.
>>nyquist(H)
You can see the characteristics of the drawn plot by right-clicking the plot and select the
characteristics (Fig. 2.27).
You can add the grids to the drawn graph by right-clicking on the graph and click the
Grid in the appeared list (Fig. 2.28). The result is shown in Fig. 2.29. Another way of adding
grid to your graph is typing the >>grid on in the MATLAB command prompt.
Figure 2.26: The Nyquist plot of the H.

2.6. DRAWING THE NYQUIST DIAGRAM OF A SYSTEM 83
Figure 2.27: Obtaining the characteristics of the Nyquist plot.

Figure 2.28: Click the grid in order to add the axes grids to the drawn Nyquist plot.
2.7. ROOT LOCUS PLOT 85
Figure 2.29: Nyquist plot of H with grid axes.
2.7 ROOT LOCUS PLOT

You can draw the root locus plot of a dynamical system with the aid of rlocus command
2 C5sC1
(Fig. 2.30). For instance, you can draw the root locus of H.s/ D 2s
s 2 C2sC3
with the aid of fol-
lowing commands:
>> H=tf([2 5 1],[1 2 3]);
>> rlocus(H), grid on
Click the Data Cursor icon in order to add a cursor to the plot. Move the added cursor;
the required gain is shown for each point (Fig. 2.31).
2s 2 C5sC1
Figure 2.30: Root locus of H.s/ D s 2 C2sC3
.
2.7. ROOT LOCUS PLOT 87
Figure 2.31: You can use the cursor to see the required gain. According to the shown box,
gain of 0.125 places one of the closed loop poles at 1:05 C 1:15i . You can use the command
pole(feedback(0.125*H,1) to check it (H=tf([2 5 1],[1 2 3])).
There is another command to draw the root locus plot: rltool. With the aid of rltool
you can see the movement of all the poles at the same time. Type the following commands:
>> H=tf([2 5 1],[1 2 3]);
>> rltool(H), grid on
The window shown in Fig. 2.32 will appear.

Click the Root Locus Editor for Loop Transfer. The poles are shown with pink red points
(Fig. 2.33).
Figure 2.32: Control system designer window will appear after you run the rltool command.
2.8. CONNECTING THE SYSTEMS TOGETHER 89
Figure 2.33: Click the root locus editor for loop transfer tab to see the root locus plot. Closed
loop poles are shown with black arrows.
Move the closed-loop poles to desired location. To do this, left-click the pink points and
move the points to the desired location without releasing the left mouse button. You can see the
movement of all points at the same time. Required gain and other useful information are shown
in the bottom of the window (Fig. 2.34).
2.8 CONNECTING THE SYSTEMS TOGETHER

Consider a block diagram such as the one shown in Fig. 2.35.
The closed-loop transfer function can be calculated with the aid of following commands:
>>C=tf([1.13 0.45],[1 0]);
>>P=tf([1],[1 3 3 1]);
Figure 2.34: Moving the poles to desired location.
1.13 + .45 1
+_ 3+3 2+3 +1
Controller Plant
Figure 2.35: Sample feedback system.
>>CL=feedback(C*P,1)
You can see the step response of closed-loop system with the >>step(feedback(C*P,1)) com-
mand.
Consider a block diagram such as the one shown in Fig. 2.36.
In this case, the closed-loop transfer function can be calculated as follows:
2.8. CONNECTING THE SYSTEMS TOGETHER 91
1
+_ 3+3 2+3 +1
1.13s + .45
Figure 2.36: Sample feedback system.
>>C=tf([1.13 0.45],[1 0]);
>>P=tf([1],[1 3 3 1]);
>>CL=feedback(P,C)
Consider the block diagram shown in Fig. 2.37 (the feedback is positive in this block
diagram).
1
++ 3+3 2+3 +1
1.13s + .45
Figure 2.37: Sample feedback system. Note that feedback is positive.
In this case, the closed-loop transfer function can be calculated as follows:
>>C=tf([1.13 0.45],[1 0]);
>>P=tf([1],[1 3 3 1]);
>>CL=feedback(P,C,+1)
Consider the connection shown in Fig. 2.38.

The transfer function of connection shown in Fig. 2.38 can be obtained with the aid of
series command, i.e., >> series(C,P).
The transfer function of connection shown in Fig. 2.39 can be obtained with the aid of
parallel command, i.e., >>parallel(C,P).
1.13 + .45 1
3 + 32 + 3 +1
Controller Plant
Figure 2.38: Series connection of systems.
1
++
3 + 32 + 3 +1
1.13 + .45

Figure 2.39: Parallel connection of systems.

93
CHAPTER 3
Introduction to Simulink®
3.1 INTRODUCTION
Simulink® is a graphical programming environment for modeling, simulating, and analyzing
multidomain dynamical systems. Using Simulink® , the user can simulate the system behavior
without using any code! The user only drag and drop the required blocks from the library to the
simulation file and connect them. This chapter introduces this important tool with the Aid of
some examples.
3.2 RUNNING THE Simulink®

In order to run the Simulink® , simply type the following command in the command window:
>>simulink
You can even click the Simulink® icon in the menu bar in order to run Simulink® (Fig. 3.1).
Figure 3.1: The Simulink® icon.
Select the Blank Model (Fig. 3.2).

Simulink® opens a new empty simulation file for you (Fig. 3.3).
Simulink® Library Browser contains the blocks which are used during the simulation.
After the user finds the desired block inside the Simulink® Library Browser, drag and drop the
block to the simulation file (Figs. 3.4 and 3.5).
Click the icon shown in Fig. 3.6 to open a new simulation file (if needed). You can press
Ctrl+N as well.
A new simulation file will be opened (Fig. 3.7).
94 3. INTRODUCTION TO Simulink®
Figure 3.2: Simulink’s first window.

3.2. RUNNING THE Simulink® 95
Figure 3.3: Simulink® environment.

Figure 3.4: Simulink® library browser.

3.2. RUNNING THE Simulink® 97
Figure 3.5: Use the shown icon if Simulink® doesn’t show the library browser.
Figure 3.6: Opening a new simulation file.

Figure 3.7: Simulation file. Blocks are placed here.
3.3 SEARCHING FOR BLOCKS

You can use the Enter search term box (Fig. 3.8) to search for a block if you don’t know where
it is. Just write what you look for and press the Enter key.
3.4 EXAMPLE 1: Simulink® OF A RLC CIRCUIT

Consider a simple RLC circuit such as the one shown in Fig. 3.9.
The key is closed at t D 0. All the initial conditions are zero, i.e., iL .0/ D 0 and vC .0/ D 0.
We want to study the circuit current.
According to Kirchhoff voltage law,
d 2 i.t/ d i.t / 1 d i.0/ E

L 2
CR C i .t / D 0 i .0/ D 0 D : (3.1)
dt dt C dt L
3.4. EXAMPLE 1: Simulink® OF A RLC CIRCUIT 99
Figure 3.8: Use the Enter search term to find the desired blocks.
S R = 1Ω L = 1H
C = 1F
V
Figure 3.9: Simple RLC circuit.
We want to simulate this Differential Equation in Simulink® . To do this, we rearange the ob-
tained equation as
d 2 i .t/ R d i .t / 1 d i .0/ E
D i .t/ i .0/ D 0 D : (3.2)
dt 2 L dt LC dt L
Use the Simulink® Library Browser to add two integrators to the simulation file. Connect them
as shown in Fig. 3.10.
2!/ "^2 i/ " !
> 1 1 >
# #
Integrator 1 Integrator
Figure 3.10: Two integrators are place in the simulation file.
Add two gain blocks, a summer and a scope to the simulation file, and connect them as
shown in Fig. 3.11.
2 /!^2 i/!
+ 1 1
+ " "
Integrator1 Integrator Scope
Gain1
-R/L
Gain
-1/L/C
Figure 3.11: Simulation diagram of Eq. (3.2).
Double-click on the gain blocks and set their Gain box (see Figs. 3.11 and 3.12) to R=L
and 1=L=C . We will define the value of R, L, and C later.
Simulink® has different solvers to solve the systems dynamic equation. The user can set
his/her desired solver. To set the solver, click on the gearbox icon (Fig. 3.13).
Select the desired solver from Solver: drop down list (Fig. 3.14).
ode45 is a quite good solver for this example. Simulink® shows the active solver, i.e., the
one which will be used during the simulation, in the bottom right side (Fig. 3.15).
We filled the values of gain blocks with R=L and 1=L=C . So, we must define the used
variables before running the simulation. Otherwise the simulation will fail and an error message
will be produced.
To define the values of aforementioned variables, go to command prompt and enter the
commands shown in Fig. 3.16.
i .0/
The differential equation has the i .0/ D 0; d dt DE
L
set of initial conditions. To set the
initial conditions, double-click the integrators and set the Initial condition box (Fig. 3.17). Set
the initial condition of left integrator to E=L and the right one to 0 (Fig. 3.18).
Figure 3.12: Setting the gain box.
Figure 3.13: Model configuration parameters icon.

Figure 3.14: Selection of solver.
Figure 3.15: Active solver is shown in the right button of the screen.
Figure 3.16: Defining the R; L; C , and E variables.

Figure 3.17: Setting the initial conditions.

E/L 0
2 /!^2 i/!
+ 1 1
+ " "
Gain1
-R/L
Gain
-1/L/C
Figure 3.18: Set the initial condition of left integrator to E=L and the right one to 0.
Enter the desired simulation interval inside the following box (see Fig. 3.19). The box is
filled with 10, so the system response is calculated for an interval of 10 s.
Figure 3.19: Setting the simulation interval.
Press the F5 button on your keyboard or click the shown icon to run the simulation
(Fig. 3.20).
Figure 3.20: Run simulation icon.
After simulation is finished, double-click the scope block to see the current (Fig. 3.21).
3.5. EXPORTING THE RESULTS TO WORKSPACE 105
Figure 3.21: Simulation result (circuit current).
3.5 EXPORTING THE RESULTS TO WORKSPACE

You can export the simulation results to workspace for further process. To do this, add a To
Workspace block to your simulation file and connect it to the rest of system as shown in Fig. 3.22.
Double-click the To Workspace block and set the Save format box to Structure With
Time (Fig. 3.23).
Run the simulation and after the simulation is finished come back to MATLAB® envi-
ronment. As shown in Fig. 3.24, two variables are added to workspace: simout and tout.
You can draw the plot of circuit current with the aid of following command. The result is
shown in Fig. 3.25.
>>plot(simout.time,simout.signals.values(:),'k'),grid minor, hold on

simout
To Workspace
2 /!^2 i/!
+ 1 1
+ " "
Gain1
-R/L
Gain
-1/L/C
Figure 3.22: Addition of to workspace block to the simulation file.
Figure 3.23: Selection of structure with time.

3.5. EXPORTING THE RESULTS TO WORKSPACE 107
Figure 3.24: Simout and tout are added to the workspace variables.
Figure 3.25: Drawing the imported data (inductor current).

3.6 EXAMPLE 2: SIMULATION IN TRANSFER FUNCTIONS
You can do the simulations contain transfer function blocks.
According to Fig. 3.9, the transfer function between the capacitor voltage and input source
is:
1
VC .s/ Cs 1
H .s/ D D 1
D 2
: (3.3)
E .s/ R C Ls C Cs LCs C RCs C 1
We want to study the capacitor voltage in presence of a ramp voltage. We draw the sim-
ulation diagram shown in Fig. 3.26. The simulation result is shown in Fig. 3.27.
1
* .!2 + #* .! + 1
Ramp Transfer Fcn Scope
Figure 3.26: Simulation block diagram.
Figure 3.27: Simulation result.

3.7. EXAMPLE 3: SIMULATION OF ELECTRIC CIRCUITS 109
3.7 EXAMPLE 3: SIMULATION OF ELECTRIC CIRCUITS
WITH THE AID OF SIMSCAP LIBRARY
We give the system dynamics to Simulink® in the previous examples, i.e., we do the modeling
phase by ourselves. We show how Simulink® can be used to simulate electric circuits in this ex-
ample. Simulink® has a library named Simscape (see Fig. 3.28) which contains electrical blocks,
i.e., resistors, capacitors, inductors, MOSFETs, AC/DC motors, etc. The Simscape library is
used in this example.
Figure 3.28: Simscape library.

We want to simulate the following RLC circuit and draw the graph of power drawn from
DC voltage source (Fig. 3.29). The switch is closed at t D 0 and the system initial conditions
are zero, i.e., iL .0/ D 0 and vC .0/ D 0.
S R = 1Ω L = 1H
C = 1F
V
Figure 3.29: Simple RLC circuit.
Go to Simscape>SimPowerSystems>Specialized Technology (Fig. 3.30).
Figure 3.30: Specialized technology branch of SimPowerSystems.
Click the Fundamental Blocks (Fig. 3.31).
Figure 3.31: Fundamental blocks.

Select the Elements (Fig. 3.32).
Figure 3.32: Elements branch of fundamental blocks.
Add three Series RLC branch (see Figs. 3.33 and 3.34) to the simulation file.
Double-click the Series RLC branch’s. Use the Branch type (see Fig. 3.35) to convert the
components to a resistor, an inductor, and a capacitor (see Fig. 3.36).
Double-click the components to set their values. Set the initial condition of induc-
tor/capacitor as shown in Fig. 3.37.
Go to Simscape>SimPowerSystems>Specialized Technology>Fundamental Blocks.
Click the Electrical Sources (see Fig. 3.38) and select a DC Voltage Source (see Fig. 3.39).
Drag and drop the DC source to simulation file (see Fig. 3.40).
We need voltmeter and ammeter to show the circuits voltage and currents.
Voltmeter and ammeter can be found in Simscape>SimPowerSystems>Specialized
Technology>Fundamental Blocks>Measurements (see Fig. 3.41).
Add voltmeter and ammeter to the circuit (see Figs. 3.42 and 3.43).
You can rotate a block place in the simulation file. To do this right-click the block and
select the Rotate and flip. You can rotate a block by left-clicking it and then pressing the Ctrl+R
(see Fig. 3.44).
All the simulations contains the SimPowerSystems blocks must contain a powergui block
(Fig. 3.45). Otherwise the simulation can’t be run and lead to an error message (see Fig. 3.46).
The complete simulation diagram is shown in Fig. 3.47
After running the simulation, the result shown in Fig. 3.48 is obtained.
Figure 3.33: Series RLC branch.

Figure 3.34: Addition of three series RLC branch to the simulation file.
Figure 3.35: Different available options for branch type.

Figure 3.36: Conversion of series RLC branches to resistor, inductor, and capacitor.
Figure 3.37: Setting the initial value of the inductor current.

Figure 3.38: Electrical sources’ branch of fundamental blocks.
Figure 3.39: DC voltage source.

Figure 3.40: Addition of DC voltage source to the simulation file.
Figure 3.41: Measurements branch of fundamental blocks.

Figure 3.42: Ammeter and voltmeter blocks.
Figure 3.43: Addition of ammeter and voltmeter to the simulation file.

Figure 3.44: Rotate and flip options.

Figure 3.45: Powergui block.
Figure 3.46: An error message occurs when you don’t add the powergui to your simulation file.
Continuous
powergui
+ i
− + + +
Current Measurement
Series RLC Branch Series RLC Branch 1 Series RLC Branch 2
+ +v ×
DC Voltage Source −
Voltage Measurement Product Scope
Figure 3.47: Addition of powergui to the simulation file.
Figure 3.48: Simulation result (DC source power).

3.8. EXAMPLE 4: SIMULATION OF NONLINEAR SYSTEMS 123
3.8 EXAMPLE 4: SIMULATION OF NONLINEAR SYSTEMS
We simulate a nonlinear system in this example. Consider the following system:
(
d
i D vLin .1 u/ vLc
dt L
(3.4)
d
v
dt C
D 1
v
RC c
C .1 u/ iCL
iL and vC are system states and u is the system input. u only takes two values: 0 and 1, i.e.,
u 2 f0; 1g. All the initial conditions are zero, i.e., iL .0/ D 0 and vC .0/ D 0.
When u D 1, (
d
i D vLin
dt L
d 1
(3.5)
v D RC
dt C
vc :
When u D 0, (
d vin vc
i
dt L
D L
d 1 1
(3.6)
v
dt C
D i
C L
v :
RC c
The simulation diagram of the system is shown in Fig. 3.49.

We want to stimulate the system with the periodic signal that is shown in Fig. 3.50. Such
a signal is produced with the aid of Pulse Generator block.
Figure 3.51 shows the simulation results.
1 iL
E 1/L +
+ s
Constant 2 Gain Integrator
0
Constant1 >0
-1/L
Gain3 Switch
vc
+ 1
Pulse -1/R/C + s
Generator
Integrator1
Gain2
Constant >0
1/C
Gain1 Switch1
Figure 3.49: Simulation diagram of Example 4 in Section 3.8.
0 T/2 T t[s]
1

Figure 3.50: Excitation signal T D 25000
D 40 s .
3.8. EXAMPLE 4: SIMULATION OF NONLINEAR SYSTEMS 125
iL
20
15
10
iL
0
-5
-10
-15
0 2 4 6 8 10 12 14 16 18 20
Time (ms)
vC
50
45
40
35
30
vC
25
20
15
10
0
0 2 4 6 8 10 12 14 16 18 20
Time (ms)
Figure 3.51: Simulation results.

3.9 SUBSYSTEM
Subsystems helps the user to produce more understandable simulation diagrams. It hides the
unnecessary details inside it.
Consider a simulation diagram such as the one shown in Fig. 3.52.
1

Step Integrator Scope
Figure 3.52: A simple simulation diagram.

Left-click on an empty point of the simulation file, don’t release the left mouse button
and draw a rectangle around the blocks. Simulink® draws a blue rectangle around the blocks
(see Fig. 3.53).
1

Figure 3.53: Draw a rectangle around the blocks.
Click the Create Subsystem (see Fig. 3.54).
1

Figure 3.54: Create subsystem icon.

After you clicked the Create Subsystem icon, Simulink® makes the subsystem. Click on
the word “Subsystem” to change it. Use descriptive names which shows the purpose of subsystem
(see Fig. 3.55).
Double-click the subsystem. Simulink® opens the subsystem for you. Remove the Step
and Scope blocks (click on the block and press the Delete key on keyboard) and replace them
with In/Out blocks. In/Out blocks can be found in the Commonly Used Blocks tab (see
Fig. 3.56).
3.9. SUBSYSTEM 127
Subsystem
Figure 3.55: Simulink® convert the selected blocks into a subsystem once create subsystem is
clicked.
Figure 3.56: In/Out blocks.

The subsystems looks like the one shown in Fig. 3.57. It has input/output port. You can
connect the subsystem to rest of system using it input/output ports.
In1 Out1
Subsystem
Figure 3.57: Simulink® adds the input/output port to the subsystem.
Connect a Step and Scope blocks to the subsystems, as shown in Fig. 3.58.
In1 Out1
Step Scope
Subsystem
Figure 3.58: Addition of step and scope block to the simulation file.
3.10 FUNCTION BLOCK

You can use the function block to produce more understandable simulation diagrams. Study the
following example to see how the function block can be used in your simulations.
3.11 EXAMPLE 5: USE OF FUNCTION BLOCK IN THE

Simulink® SIMULATIONS
Consider a dynamical system with the equation dy.t dt
/
D 3 y .t/ C x.t/2 . x.t / and y.t/ show
the system input and output, respectively. We want to simulate the system with a sinusoidal
input, x .t/ D sin.t /. The initial condition is zero (y .0/ D 0/.
Solution:
1st way:
You can draw the following simulation diagram (see Fig. 3.59).
2nd way:
You can use the function block to simulate the system. Function block can be found at
SIMULINK>User Defined Functions>Fcn (see Fig. 3.60).
3.11. EXAMPLE 5: USE OF FUNCTION BLOCK IN THE Simulink® SIMULATIONS 129
_ !"
1
% +
+ #
Sine Wave Product Integrator Scope
Gain
3
Figure 3.59: Simulink® diagram of Example 5 in Section 3.11. No function block is used in this
simulation.
Figure 3.60: Function block.

Add a Mux block (see Fig. 3.61) to the simulation file and connect it to the Fcn block
(Fig. 3.62).
Figure 3.61: Multiplexer block.
f(u)
Fcn
Figure 3.62: Multiplexer block is connected to the input of function block.
Double-click the Fcn block and fill the Expression box, as shown in Fig. 3.63.
3.11. EXAMPLE 5: USE OF FUNCTION BLOCK IN THE Simulink® SIMULATIONS 131
Figure 3.63: Filling the expression box of function block.
The u.1/ and u.2/ are the inputs of the multiplexer block. The u.1/ is the topmost input
(see Fig. 3.64).
u(1)
f(u)
u(2)
Fcn
Figure 3.64: Input of function block are u.1/ and u.2/. u.1/ is the topmost input.
Figure 3.65 shows the simulation diagram of the system.
1
f(u)
Sine Wave
Fcn Scope
Figure 3.65: Simulink® diagram of Example 5 in Section 3.11. Function block is used in this
simulation.
Use of function blocks (when possible) is suggested. The simulation result is shown in
Fig. 3.66.
3.12 SIMULATION OF DISCRETE TIME SYSTEMS

We want to simulate a discrete time system with the following equation:
2 x .n 1/ C 6 x .n 2/ D x .n/ ; x . 1/ D 11x . 2/ D 7:
We want to obtain x .n/ ; n 0. Some of the first terms can be calculated manually
x .0/ D 2 11 C 6 7 D 64
x .1/ D 2 64 C 6 11 D 194
x .2/ D 2 194 C 6 64 D 772:
We use the Delay block (Fig. 3.67) to simulate the system.

Add two delay blocks to your simulation and connect them together, as shown in Fig. 3.68.
3.12. SIMULATION OF DISCRETE TIME SYSTEMS 133
Figure 3.67: Delay block.
The system can be simulated with the aid of simulation diagram shown in Fig. 3.69.
The initial condition can be set with the aid of Initial condition box. Double-click the delay
blocks (Figs. 3.70 and 3.71), after the block is opened; set their corresponding initial condition
value (Fig. 3.72).
(") ("-1) ("-2)

-1 -1
Delay1 Delay
Figure 3.68: Two delay blocks are added to the simulation file.
Scope
+
+ -1 -1
Delay1 Delay
Gain
Gain1
Figure 3.69: Completed simulation file.

Figure 3.70: Setting the delay1.

Figure 3.71: Setting the delay.

Scope
11 7
+
+ -1 -1
Delay1 Delay
Gain
Gain1
Figure 3.72: Initial value of left delay block is set to 11 and the right block is set to 7.
Click the gear icon and set the Solver to Discrete (no continuous state). See Figs. 3.73
and 3.74.
Figure 3.73: Model configuration parameters icon.
Simulation result is shown in Fig. 3.75.

Figure 3.74: Setting the solver type to discrete (no continuous states).
3.13. EXAMPLE 6: SIMULATION OF DISCRETE TIME SYSTEMS 139
3.13 EXAMPLE 6: SIMULATION OF DISCRETE TIME

SYSTEMS
We want to simulate another system. Consider the following system:
y .n/ 3 y .n 1/ D 4n 1
; y . 1/ D 0:
We write the system as:
y .n/ D 3 y .n 1/ C 4n 1
; y . 1/ D 0:
The first few terms can be calculated:
y .0/ D 3 0 C 4 1 D 0:25;
3 7
y .1/ D C 1 D D 1:75;
4 4
21 37
y .2/ D C4D D 9:25:
4 4
The system can be simulated with the aid of following simulation diagram (see Fig. 3.76).
Scope
f(u) + -1
+
Clock Zero-Order Fcn Delay
Hold
Gain
1
Figure 3.76: Simulation diagram of Example 6 in Section 3.13.
In order to produce the excitation term (4n 1 term), we used a Clock block, a Zero-Order
Hold block, and a function block (Fig. 3.77). The clock block produces a continuous output.
Zero-Order Hold converts the output of the clock block to a staircase signal (see Figs. 3.78,
3.79, and 3.80).
Figure 3.77: Setting the function block.

Figure 3.78: The clock block.

Figure 3.79: The zero-order hold block.

Figure 3.80: The delay block setting.
3.14 EXAMPLE 7: SIMULATION OF DISCRETE TIME

SYSTEMS
We want to simulate the following system in this example:
1 1 1
y .n/ y .n 1/ D x .n/ C x .n 1/ ; y . 1/ D 0;
2 4 4
y .n/ is the system output and x .n/ is the input. x.n/ is unit step function,

0 n<0
x .n/ D
1 n 0:
The given equation can be rearranged as:
1 1 1
y .n/ D y .n 1/ C x .n/ C x .n 1/ ; y . 1/ D 0:
2 4 4
3.15. SHOWING TWO OR MORE SIGNALS SIMULTANEOUSLY 145
The first few terms can be calculated as:
1
y .0/ D D 0:25;
4
5
y .1/ D D 0:625;
8
13
y .2/ D D 0:812;
16
29
y .3/ D D 0:906:
32
The simulation diagram of the system is shown in Fig. 3.81. The used blocks settings are
given in Figs. 3.82–3.84.
0.25 Scope
Gain1
+
+
-1
+
1 -1 0.25
Delay
Constant Delay1 Gain2
Gain
0.5
Figure 3.81: The simulation diagram of Example 7 in Section 3.14.
3.15 SHOWING TWO OR MORE SIGNALS

SIMULTANEOUSLY
You can show two or more signals simultaneously with the aid of Mux block (see Fig. 3.85).
In the following simulation file, three different signals are shown on an Scope. Simulation
result is shown in Fig. 3.86. Multiplexer settings are shown in Fig. 3.87 and the simulation result
is shown in Fig. 3.88.
Figure 3.82: Setting of the constant block.

Figure 3.83: Setting of the delay block.

Figure 3.84: Setting of the delay1 block.

Figure 3.85: Multiplexer block.

Sine Wave
Scope
Step
Repeating
Sequence
Stair
Figure 3.86: Showing three different signals on the same scope.
Figure 3.87: Setting of the multiplexer block.

3.16. SIMULATION OF A CLOSED-LOOP CONTROL SYSTEM IN Simulink® 151
3.16 SIMULATION OF A CLOSED-LOOP CONTROL

SYSTEM IN Simulink®
We want to simulate the closed-loop control system shown in Fig. 3.89.
1.13s + .45 1
+_
s s3 + 3s2 + 3s +1
Controller Plant
Figure 3.89: Sample closed-loop system.

Add two Transfer Function block to the simulation file (Figs. 3.90 and 3.91).
Figure 3.90: Transfer function block.
1 1
s+1 s+1
Transfer Fcn 1 Transfer Fcn
Figure 3.91: Adding two transfer function block to the simulation file.
Add a summer to the simulation file (Figs. 3.92 and 3.93).
Figure 3.92: The summer block.
1 1
+
+ s+1 s+1
Figure 3.93: Addition of summer block to the simulation file.

Double-click on the summer and set the list of signs to C (Figs. 3.94 and 3.95).
Figure 3.94: Set the List of signs to C in order to simulate negative feedback.
1 1
+
− +1 +1
Figure 3.95: The summer can simulate negative feedback now.

Connect the blocks together (Fig. 3.96). Click on the “Transfer FC1” and “Transfer Fcn”
labels and change them to “Controller” and “Plant,” respectively. This have no effect on the
simulation but it makes the simulation diagram more meaningful from the user point of view.
1 1
+
− +1 +1
Controller Plant
Figure 3.96: Connecting the blocks together.
Double-click on the Controller and Plant and set the transfer functions, as shown in
Figs. 3.97 and 3.98. The simulation diagram looks like the one shown in Fig. 3.99 after you
change the coefficients.
You can make the Plant a little bit bigger to see the equation of transfer function inside it
(Fig. 3.100).
Add a step block to the simulation file (Figs. 3.101 and 3.102).
Double-click the simulation file. Step time is set to 1 by default (Fig. 3.103). This cause
the block to produce the signal shown in Fig. 3.104. Turn the Step time to 0 if you want to
stimulate the system with the one shown in Fig. 3.105.
Add a Scope block to the system (Fig. 3.106). An scope block is added to see the plant
output (Fig. 3.107). Run the simulation with the aid of the button shown (Fig. 3.108). See the
results in Fig. 3.109.
Figure 3.97: Entering the coefficients of controller.

Figure 3.98: Entering the coefficients of plant.
1.13s + .45 1
+
− s den(s)
Controller Plant
Figure 3.99: The simulation file after new coefficients entered.

1.13 + .45 1
+
− 3 + 32 + 3 + 1
Controller Plant
Figure 3.100: Drag the plant sides in order to show the transfer function inside it.
Figure 3.101: The step block.

1.13 + .45 1
+
− 3 + 32 + 3 + 1
Step Controller Plant
Figure 3.102: Completed simulation file.
Figure 3.103: Step blocks setting.

1.5
0.5
-0.5
-2 -1 0 1 2
Figure 3.104: Output of step block when step time D 1.
1.5
0.5
-0.5
-2 -1 0 1 2
Figure 3.105: Output of step block when step time D 0.

Figure 3.106: Scope block.
1.13 + .45 1
+
− 3 + 32 + 3 + 1
Step Controller Plant Scope
Figure 3.107: An scope block is added to see the plant output.

Figure 3.108: Run the simulation with the aid of the button shown.

163
CHAPTER 4
Controller Design in
MATLAB®
4.1 INTRODUCTION
MATLAB® can be used to design different types of controllers. This chapter introduces some
of the most important tools provided by MATLAB® to design linear controllers.
PID controllers are the most commonly used controller used in real world. About 90% of
controllers used in industry are PID controllers. Fortunatly, MATLAB® has powerful functions
to tune this important controller. Using MATLAB® a PID controller can be tuned within few
minutes!
4.2 PID CONTROLLER DESIGN IN MATLAB®

MATLAB® has a powerfull function named pidTuner to design PID controllers. Its syntax is:
pidTuner.plantName/
We show how to use it, with the aid of an example.

We want to design a PID controller for
sC1 sC1
H .s/ D D 3 2
:
.s C 2/ .s C 3/ .s C 4/ s C 9s C 26s C 24
We must define the dynamical equation in the MATLAB® environment. We do the job with
the aid of one of the following commands.
>>H=tf([1 1],[1 9 26 24]);
or
>>H=zpk([-1],[-2 -3 -4],1);
The step function of this transfer function is shown in Fig. 4.1. So, there is a need to a
controller, since it cannot track a step signal, i.e., the steady-state error is not zero. The systems
is slow as well.
164 4. CONTROLLER DESIGN IN MATLAB®
Step Response
0.05
0.045
0.04
0.035
0.03
Amplitude
0.025
0.02
0.015
0.01
0.005
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time (s)
Figure 4.1: Step response of H.
We type the following command in MATLAB® command prompt:
>>pidTuner(H)
This opens the PID Tuner window (Fig. 4.2).

Select the controller type first. Available options are shown in Fig. 4.3. We select a PI for
this application (we want to obtain zero steady-state error so we need the integrator term).
Use the sliders to obtain your desired response (Fig. 4.4).
You can do the tuning in the frequency domain if you prefer. If you like to do so, set the
Domain to Frequency (Fig. 4.5).
When you switch to frequency domain, your slider changes to Bandwidth and Phase
Margin, as shown in Fig. 4.6.
Change the sliders to obtain the desired response. The controller parameters, i.e., the PI
controller coefficients are shown in the bottom right side (Fig. 4.7).
4.2. PID CONTROLLER DESIGN IN MATLAB® 165
Figure 4.2: PID tuner window.

Figure 4.3: Different available types of controllers.
Figure 4.4: Move the sliders to tune the controller parameters.
Figure 4.5: You can tune the controller in the frequency domain.
Figure 4.6: Tuning in the frequency domain.

Figure 4.7: The controller parameters are shown in the bottom left side of the window.
Always ensure that the control signal (see Fig. 4.8) is within the desired range. Obtaining
a desired output response (Fig. 4.7) is not enough; you must ensure the controller is able to
produce the control signal as well. To ensure that control signals are withing their range, click
the Add Plot icon and select the Controller Effort (Fig. 4.9).
Control Signal
+
−
Controller Plant Output
Figure 4.8: Control signal (control effort) is the signal produced by the controller.
The controller effort window will be added to the PID Tuner window (Fig. 4.10).
Figure 4.9: Plot of controller effort can be added to the main window by clicking the controller
effort.
Figure 4.10: The designer can see the controller effort while he/she changes the sliders.
If the control signal is too large, repeat the design process, i.e., change the sliders until you
obtain a more suitable response. Openning the controller effort plot is always suggested. Since
the designer can see the control signal during the design process.
We want to regenerate the graphics shown in Fig. 4.10. Since the Kp D 41:59 and Ki D
105:1 the controller equation is Hc .s/ D 41:59 C 105:1 s
. The following code (Fig. 4.11) draws
the closed-loop step response. The result is shown in Fig. 4.12 and is the same with the graph
shown in bottom left of the window (see Fig. 4.10).
The following code (Fig. 4.13) draws the control signal (control effort) for a step input.
The result is shown in Fig. 4.14 and is the same with the graph shown in bottom right of the
window (see Fig. 4.10).
Use the Export icon to export the designed controller into the MATLAB® workspace
(Fig. 4.15).
Figure 4.11: Drawing the step response of the closed-loop system.
Figure 4.12: Step response of the closed-loop system.

Figure 4.13: Drawing the control effort for a step reference.
Figure 4.14: Control effort signal for a step reference.

Figure 4.15: Export icon.
After you click the Export icon, the following window will appear (Fig. 4.16). Click
OK to export the controller. The designed controller is saved in a variable named C inside the
workspace.
Figure 4.16: The designed controller is exported to Workspace.
You can see the closed-loop step response with the aid of following command:
>>step(feedback(H*C,1))
4.3 TUNING THE PID CONTROLLERS INSIDE THE

Simulink® ENVIRONMENT
You can do the tuning of PID controller inside the Simulink® environment. For example, con-
sider the system shown in Fig. 4.17.
Double-click the PID Controller block (Fig. 4.18).
4.3. TUNING THE PID CONTROLLERS INSIDE THE Simulink® ENVIRONMENT 173
+ PID(s) +1
− 3 + 92 + 26 + 24
Step PID Controller Transfer Fcn Output
Figure 4.17: The Simulink® diagram of a closed-loop system.
Figure 4.18: PID Controller settings.

Select your desired type of controller using the Controller drop-down list (Fig. 4.19).
Figure 4.19: Select the desired type of controller from the drop-down list.
Click the Tune... button to start the tuning (Fig. 4.20).
Figure 4.20: Click the Tune…button to start the tuning.
The Simulink® runs the PID Tuner for you. You can tune the block in the same way
studied before (Fig. 4.21). Once you obtain the desired output response (with acceptable control
4.3. TUNING THE PID CONTROLLERS INSIDE THE Simulink® ENVIRONMENT 175
effort) you can click the Update Block icon and Simulink® update the PID block with the new
coefficients (Fig. 4.22).
Figure 4.21: Simulink® runs the PID tuner after you clicked the tune ….
Figure 4.22: Click the update block to update the controller coefficients.
4.4 DESIGN OF LINEAR CONTROLLERS WITH

SISOTOOL (CONTROL SYSTEM DESIGNER)
MATLAB® has a application named Control System Designer which can be used for design
of linear control systems. You can use the commands shown in Fig. 4.23 to enter the Control
System Designer environment. The Control System Designer environment is shown in Fig. 4.24.
Figure 4.23: Control system designer can be run with the command controlSystemDesigner.
We want to design a linear controller for H.
You can select the desired structure using the “Edit Architecture” button as shown in
Fig. 4.25. Using the icons you can import the transfer functions from the Workspace.
You can use the “Tuning Methods” button (Fig. 4.26) and “Automated Tuning” section
to automatically tune the controller (Block C in Fig. 4.25). If you click the “PID Tuning,” a
window like that shown in Fig. 4.27 will appear.
You can do the tuning either in time domain or frequency domain (Fig. 4.28).
4.4. DESIGN OF LINEAR CONTROLLERS WITH SISOTOOL 177
Figure 4.24: “Control system designer” window.

Figure 4.25: Selecting the desired control topology. The one shown with black arrow is the most
commonly used topology.
Figure 4.26: “Tuning methods” button.

Figure 4.27: PID tuning window.

Figure 4.28: Tuning can be done either in time domain or frequency domain.
Figure 4.29: Tuning a PI controller is an iterative process.
To tune the controller (Fig. 4.29), do the following.
1. Select the desired controller type (P, PI, PID, etc.). Always start with simple P-type con-
troller. If response is not acceptable then use PI or PID.
2. Select the desired domain (time or frequency).
3. Move the sliders.
4. Click the Update Compensator Button.
5. See the step response and decide whether the response is good or not. If you are not happy,
redo the previous steps, i.e., move the sliders again.
You can right-click on the shown step response and use the Characteristics in order to see
the characteristics of the response (Fig. 4.30).
Figure 4.30: Reading the step response characteristics.
You can export the tuned controller to Workspace. To do this, click the Export button
(Fig. 4.31).
Figure 4.31: Exporting the controller to workspace.
A window like the one shown in Fig. 4.32 will appear. Check the box to the left of C and
click the Export button.
A new variable named C will be added to Workspace after you clicked Export button. It
contains the designed controller. You can see the controller equation with the aid of tf command
(Fig. 4.33).
Figure 4.32: Export model window.
Figure 4.33: Equation of designed controller.

4.5. LOOP SHAPING 185
4.5 LOOP SHAPING
You can use design the controller by using the loop shaping method (Fig. 4.34).
Figure 4.34: Loop Shaping button.

Assume a control structure like that shown in Fig. 4.35. C(s) and P(s) show controller and
plant transfer function, respectively.
U(s) R(s)
+
−
C(s) P(s)
Figure 4.35: Schematic of a simple closed loop control system. C(s) and P(s) show the controller
and plant, respectively.
The closed-loop transfer function is:
R.s/ C .s/ P .s/

HCL .s/ D D :
U.s/ 1 C C .s/ P .s/
K
The product of C .s/ P .s/ is called “loop transfer function.” Assume that C .s/ P .s/ D s
. In
this case the closed-loop transfer function is
K
R.s/ s K
HCL .s/ D D K
D :
U.s/ 1C s s C K
If we apply a step function to such a system, it tracks the step function with zero steady-state
error within K5 s. For example, if C .s/ P .s/ D 1000
s
, it takes about 1000 5
D 5 ms, to track the
step command. The response is shown in Fig. 4.36. Note that the response has no overshoot or
oscillatory nature.
!n 2
As another example, assume that C .s/ P .s/ D s.sC2 !n /
. In this case, the closed-loop
transfer function is
!n 2
R.s/ s.sC2 !n / !n 2
HCL .s/ D D D :
U.s/ !n 2
1 C s.sC2! s 2 C 2!n s C !n 2
n /
Based on the values of and !n , the response takes different forms. Figure 4.37, shows the step
response for two different values of . As you see, an increase in decrease the oscillatory nature.
Loop shaping takes the desired loop transfer function (i.e., Hdesired .s/) and design the
controller block (block C in Fig. 4.35) such that the error between Hdesired .s/ C.s/P .s/ is
minimized.
One may ask, C.s/ D Hdesired
P .s/
.s/
makes the error zero since,
Hdesired .s/
Hdesired .s/ C.s/P .s/ D Hdesired .s/ P .s/ D 0:
P .s/
Step Response
1
0.9
0.8
0.7
0.6
Amplitude
0.5
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9
Time (s) ×10-3
1000
Figure 4.36: Step response of sC1000
.
Step Response
1.8
ζ = 0.7
1.6 ζ = 0.1
1.4
1.2
Amplitude
1
0.8
0.6
0.4
0.2
0
0 0.01 0.02 0.03 0.04 0.05 0.06
Time (s)
Figure 4.37: Effect of damping ratio ( ) on step response of a second-order system.
Although the approach seems correct, it is not applicable except for simple cases. Studying an
4
example is quite usefull. For example, assume that P .s/ D .sC7/.sC8/ and Hdesired .s/ D 10
s
. So,
C.s/ D Hdesired
P .s/
.s/
D 10.sC7/.sC8/
4s
. The obtained controller is not proper, so it is not realizable.
The Loop Shaping button in “AUTOMATED TUNING” section of Fig. 4.34, uses op-
timization techniques to obtain the proper controller C(s). User translates the requirement into
the Hdesired .s/ for example if one need a response free of overshoot with settling time less than
5 ms, he/she may use Hdesired .s/ D C.s/P .s/ D 1000s
is a good option. If an overshoot of less than
10002
5% is acceptable, one may use Hdesired .s/ D C.s/P .s/ D s.sC20:71000/ . Normally, Hdesired .s/
is selected among the first- and second-order transfer functions to avoid increase in the order of
designed controller.
When you click the Loop Shaping button in Fig. 4.34, the window shown in Fig. 4.38 will
appear. You enter the selected Hdesired .s/ into the “Target open-loop shape (LTI)” box using the
command “tf.” “Frequency range for loop shaping [wmin,wmax]:” box takes the frequency por-
tion where the optimization must be done. Normally, we need the overlap in the low frequency
range.
Figure 4.38: Loop Shaping window.
“Desired controller order” slider bar takes the controller order. You start by lower values
(i.e., 1). Design process is started by pressing the “Update Compensator” button. When the
order of controller is too low, the software shows a warning and asks you to increase the order.
In this case you increase the order by one. If the warning comes out again, you increase it again.
4.6 MANUAL CONTROLLER DESIGN
You can add your desired pole/zero to the controller. Adding the pole/zero can be done by right-
clicking one of the diagrams shown in the Control System Designer environment (Fig. 4.39).
You can add the desired pole/zero by selecting the “Add Pole/Zero” as shown in Fig. 4.40. As
shown in Fig. 4.40 you can add integrator, lead, lag, etc. easily by clicking the corresponding
choice. If the steady-state error for a step input is not zero, add an integrator to the loop.
You can push the left mouse button on the Bode diagram and (without releasing the left
mouse button) move the diagram up/down. This lets you to set the gain of system.
The P.M. and G.M. on the Bode plot shows Phase Margin and Gain Margin, respectively.
You can see the equation of designed controller by right-clicking on the plots and select Edit
Compensator (Fig. 4.41).
Figure 4.39: Right-clicking on the plot opens the menu.

4.6. MANUAL CONTROLLER DESIGN 191
Figure 4.40: “Add pole/zero” menu.
Figure 4.41: Click the edit compensator to see the equation of controller.
193
CHAPTER 5
Introduction to System
Identification Toolbox™
5.1 INTRODUCTION
In the previous chapters we assumed that the dynamical model of plant is known. If you don’t
know the dynamical model of plant in hand, you can make a model for it based on the in-
put/output data. System identification is a methodology for building mathematical models of
dynamic systems using measurements of the system’s input and output signals.
This chapter introduces the System Identification Toolbox™ with an example.
5.2 ILLUSTRATIVE EXAMPLE

Consider a system with the following dynamical equation:
.s C 1/ .s C 2/ 30s 2 C 90s C 60
H .s/ D 30 D 3 : (5.1)
.s C 3/ .s C 4/ .s C 5/ s C 12s 2 C 47s C 60
We draw the following simulation diagram (Fig. 5.1).
! "
To Workspace To Workspace1
302 + 90 + 60
+ 122 + 47 + 60
3
Step Transfer Fcn Scope
Figure 5.1: Simulink® diagram of the assumed system.
The system is stimulated with a step function. Meanwile the input and output of the
system are send to workspace for further process.
The “To Workspace” blocks Save format is set to Structure With Time (Figs. 5.2 and 5.3).
Click the gear icon. Do the setting such as the one shown in Fig. 5.4. This asks the
Simulink® to sample the input/output with the rate of 10 ms (Fig. 5.5).
194 5. INTRODUCTION TO SYSTEM IDENTIFICATION Toolbox™
Figure 5.2: To Workspace block settings.

5.2. ILLUSTRATIVE EXAMPLE 195
Figure 5.3: To Workspace1 block settings.
Figure 5.4: Model configuration parameter icon.

Figure 5.5: Selection of solver.
Run the simulation. After the simulation is done, three new variables are added to
MATLAB® Workspace. These are tout, u, and y (Fig. 5.6).
Figure 5.6: New variables (tout, u, and y) are added to the workspace.
u and y are structures. The following commands extract their data.
>>U=u.signals.values;
>>Y=y.signals.values;
The following command draws the input and output on the same graph (Fig. 5.7).
>>plot(tout,U,tout,Y),
grid minor,xlabel(`Time(s)'), title(Ìnput---output of system')

Input-Output of System
3
2.5
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
Time (s)
Figure 5.7: Input/output signals of system (Fig. 5.1) are drawn on the same graph.
Run the System Identification Toolbox™ by typing the following command.
>>ident
The result is shown in Fig. 5.8.

Click the import data>Time domain data… to enter the signals into the System Iden-
tification Toolbox™ (Fig. 5.9).
After you clicked the import data>Time domain data…, the Import Data window will
be shown. Fill the boxes as shown in Fig. 5.10 and click the Import and Close buttons, respec-
tively.
Check the Time plot to see the graph of imported data. The graph is shown in Fig. 5.11.
Click the Estimate> Transfer Function Models … in order to make a transfer function
model for the imported data (Fig. 5.12).
Figure 5.8: System identification Toolbox™ .
Figure 5.9: Importing the time domain data.

Figure 5.10: Import data window.

Figure 5.11: Time domain plot of input/output signals.

Figure 5.12: Select transfer function models… to estimate a transfer function for the in-
put/output data.
Set the Number of poles and zeros (Fig. 5.13). You must have some insight about your
system to set these values, i.e., you know the order of your system. We asked the toolbox to make
a third-order model for us. After you enter the number of polse and zeros, click the Estimate
button. The toolbox start processing (Fig. 5.14).
After calculations finished, you can see the summary of what has been done in the Result
section of the window (Figs. 5.14 and 5.15). Fit to estimation data must be a larg number (i.e.,
80%). It can be an indicator of how reliable our model is. Since we have no noise and we set the
number of poles and zeros, it reaches 100% in this example. However, in a real-world application,
i.e., when the data is collected with sensors, there are some noise and reaching 100% seems
difficult.
Figure 5.13: Defining the order of model, i.e., how many poles/zeros the model have.
Figure 5.14: Plant identification progress window shows the iterations.

Figure 5.15: Results of calculations.
Click Close to return to main window of toolbox. If you want to send the calculated trans-
fer function to MATLAB® workspace, you can drag the tf1 and release it on the To Workspace
button (Fig. 5.16).
Figure 5.16: Transfering the obtained model to Workspace.
You can see a variable named tf1 is added to the Workspace (Fig. 5.17).
Type tf1 at the Workspace to see the obtained transfer function equation (Fig. 5.18).
Figure 5.17: New variable named tf1 is added to the workspace.
Figure 5.18: The equation of estimated model.

If you change the number of poles to 2 and number of zeros to 1, the Fit to estimation
decrease a little bit (Fig. 5.19).
Figure 5.19: Estimating a second-order model to the imported data. In this case the Fit to
estimation decreased slightly.
In this case the toolbox gives the following model:

29:4s C 26:24
H2pole;1zero .s/ D : (5.2)
s2 C 9:438s C 26:51
The following code compares Bode plot of this transfer function with the original transfer func-
tion. The result is shown in Fig. 5.20. As shown, the two transfer functions overlap.
>> H=tf([30 90 60],[1 12 47 60]); %original system
>> I=tf([29.4 26.24],[1 9.438 26.51]); %H2pole,1zero
>> bode(H,I), grid on

Bode Diagram
10
0
Magnitude (dB)
-10
-20
-30
45
Phase (deg)
-45
-90
10-2 10-1 100 101 102 103
Frequency (rad/s)
Figure 5.20: Drawing of the original systems frequency response and the second-order estimate
on the same graph.
209
References
[1] Katsuhiko Ogata, Modern Control Engineering, Pearson, 2009. DOI: 10.1115/1.3426465.
[2] Benjamin C. Kuo and Farid Golnaraghi, Automatic Control Systems, Wiley, 2002.
[3] Gene F. Franklin, J. Da Powell, and Abbas Emami Naeini, Feedback Control of Dynamic
Systems, Pearson, 2014.
[4] Chi-Tsong Chen, Linear System Theory and Design, Oxford University, 2012.
[5] Katsuhiko Ogata, Matlab for Control Engineers, Pearson, 2007.
[6] Simscape User’s Guide, MathWorks, 2015.
[7] Lennart Lejung, System Identification: Theory for the User, Prentice Hall, 1999.
[8] Johan Schoukens, Rik Pintelon, and Yves Rolain, Mastering System Identification in 100
Exercises, Wiley, IEEE Press, 2012. DOI: 10.1002/9781118218532.
[9] Katsuhiko Ogata, Matlab for Control Engineers, Pearson, 2007.
[10] Brian Hahn and Daniel T. Valentine, Essential Matlab for Engineers and Scientists, Aca-
demic Press, 2016.
[11] Farzin Asadi, Computer Techniques for Dynamic Modeling for DC-DC Power Converters,
Synthesis Lectures on Power Electronics, Morgan & Claypool Publishers, 2018.
[12] Farzin Asadi, Robust Control of DC-DC Converters: The Kharitonov’s Theorem Approach with
MATLAB® Codes, Synthesis Lectures on Power Electronics, Morgan & Claypool Publish-
ers, 2018.
[13] Farzin Asadi and Key Eguchi, Dynamics and Control of DC-DC Converters, Synthesis Lec-
tures on Power Electronics, Morgan & Claypool Publishers, 2018.
[14] Farzin Asadi, Sawai Pongswadt, Kei Eguchi, and Ngo Lam Trung, Modeling Uncertainties
in DC-DC Converters with MATLAB® and PLECS® , Synthesis Lectures on Power Elec-
tronics, Morgan & Claypool Publishers, 2018.
211
Authors’ Biographies
FARZIN ASADI
Farzin Asadi is with the Department of Mechatronics En-
gineering at the Kocaeli University, Kocaeli, Turkey. Farzin
has published 30 international papers and 10 books. He is on
the editorial board of 6 scientific journals as well. His research
interests include switching converters, control theory, robust
control of power electronics converters, and robotics.
ROBERT E. BOLANOS
Robert E. Bolanos received a B.S. in Electrical Engineering
from the University of Texas at San Antonio, and an M.S.
in Electrical Engineering from the University of Idaho. Cur-
rently, he is a Principal Engineer at Southwest Research Insti-
tute, Space Science and Engineering Division 15. Mr. Bolanos
designs space-rated switch power supplies, RF oscillators, and
high voltage pulsers. He has worked on both NASA and ESA
space projects such as IBEX, JUICE, BepiColombo, and EU-
ROPA (MASPEX).
212 AUTHORS’ BIOGRAPHIES
JORGE RODRÍGUEZ
Jorge Rodríguez received a B.S. in Electronics Engineering
and a Master’s in Electronics Engineering, both at Carlos III
University of Madrid (Spain). Currently, he is working as a
R&D hardware engineer at Power Smart Control focusing on
SoC devices, developing HIL systems, and as a support engi-
neer for PSIM and SmartCtrl software.

2019 - Feedback Control Systems - Matlab - Simulink Approach - Farzin Asadi, Robert E. Bolanos, Jorge Rodriguez

Uploaded by

Copyright:

Available Formats

2019 - Feedback Control Systems - Matlab - Simulink Approach - Farzin Asadi, Robert E. Bolanos, Jorge Rodriguez

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2019 - Feedback Control Systems - Matlab - Simulink Approach - Farzin Asadi, Robert E. Bolanos, Jorge Rodriguez

Uploaded by

Copyright:

Available Formats

Feedback Control Systems

The MATLAB®/Simulink® Approach

Feedback Control Systems: The MATLAB® /Simulink® Approach

Mobile Robotics for Multidisciplinary Study

Control System Synthesis: A Factorization Approach, Part I

Control System Synthesis: A Factorization Approach, Part II

The Reaction Wheel Pendulum

Feedback Control Systems: The MATLAB® /Simulink® Approach

ISBN: 9781681735399 paperback

A Publication in the Morgan & Claypool Publishers series

SYNTHESIS LECTURES ON CONTROL AND MECHATRONICS #5

2 Commonly Used Commands in Analysis of Control Systems . . . . . . . . . . . . . . 55

4 Controller Design in MATLAB® . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

5 Introduction to System Identiﬁcation Toolbox™ . . . . . . . . . . . . . . . . . . . . . . . 193

Authors’ Biographies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

Farzin Asadi, Robert E. Bolanos, and Jorge Rodríguez

Farzin Asadi, Robert E. Bolanos, and Jorge Rodríguez

Figure 1.1: Diﬀerent parts of MATLAB® .

1.2 DIFFERENT PARTS OF MATLAB®

Figure 1.2: Command window.

Figure 1.3: Workspace.

Figure 1.4: Current folder.

Figure 1.5: Menu bar.

Figure 1.6: Simulink icon.

1.3 MATLAB’S EDITOR

The editor is shown in Fig. 1.8.

Figure 1.7: Layout icon.

Figure 1.8: MATLAB’s editor.

Figure 1.10: Run icon.

1.4 MATLAB’S HELP SYSTEM

1.5 MATLAB’S KEYWORDS

Figure 1.11: Help window.

Figure 1.13: MATLAB’s online help.

Figure 1.14: Keywords are shown in blue.

break case catch continue

1.6 MATLAB’S TOOLBOXES

Table 1.2: Some of MATLAB’s toolboxes

MATLAB Fixed-Point Designer Paralle Computing Toolbox

1.8 BASIC OPERATORS

Table 1.4: MATLAB’s basic operators

Operator Explanation Operator Explanation

0.2500 0.4000 0.5000

1.9 LOGICAL OPERATORS

1.10 TRIGONOMETRICS FUNCTIONS

Table 1.5: MATLAB’s logical operators

Operator Explanation Example

Table 1.6: MATLAB’s trigonometric functions (Continues)

Function Explanation Example

Table 1.6: (Continued) MATLAB’s trigonometric functions (Continues)

atan Calculates the arc tangent. Result is in Radyans. >>atan(1)

cotd Calculates the cotangent of an angle. Operand >>cotd(1)

1.11 LOGARITHMIC AND EXPONENTIAL FUNCTIONS

1.12 COMPLEX NUMBERS FUNCTIONS

1.13 ROUNDING FUNCTIONS

1.14 REMAINDER FUNCTIONS

Table 1.7: MATLAB’s logarithmic and exponential functions

Function Explanation Example

Table 1.8: Conversion of simple expression to MATLAB®

Table 1.9: MATLAB’s functions related to complex numbers

We enter the data to MATLAB® (Fig. 1.46). ˇ ˇ