FUZZY CONTROLLERS

RECENT ADVANCES IN
THEORY AND
APPLICATIONS

Edited by Sohail Iqbal






Fuzzy Controllers Recent Advances in Theory and Applications
http://dx.doi.org/10.5772/2622
Edited by Sohail Iqbal

Contributors
Teodor Lucian Grigorie, Ruxandra Mihaela Botez, Andrei Vladimir Popov, Stela Rusu-Anghel,
Lucian Gherman, Omer Aydogdu, Ramazan Akkaya, M.H. Fazel Zarandi, Fereidoon Moghadas
Nejad, H. Zakeri, Xin Wang, Edwin E. Yaz, James Long, Tim Miller, S. Bouallgue, J. Haggge,
M. Benrejeb, Meriem Nachidi, Ahmed El Hajjaji, Ying-Shieh Kung, Chung-Chun Huang, Liang-
Chiao Huang, Ping Zhang, Guodong Gao, Maguid H. M. Hassan, M. Chadli, A. El Hajjaji, Pedro
Ponce, Arturo Molina, Rafael Mendoza, Kwanchai Sinthipsomboon, Issaree Hunsacharoonroj,
Josept Khedari, Watcharin Po-ngaen, Pornjit Pratumsuwan, Yousif I. Al Mashhadany, Carlos
Andr Guerra Fonseca, Fbio Meneghetti Ugulino de Arajo, Marconi Cmara Rodrigues, Nora
Boumella, Juan Carlos Figueroa, Sohail Iqbal, Muhammad M.A.S. Mahmoud, Morteza Seidi,
Marzieh Hajiaghamemar, Bruce Segee, Wudhichai Assawinchaichote, Yassine Manai,
Mohamed Benrejeb, Mavungu Masiala, Mohsen Ghribi, Azeddine Kaddouri, Georgios A.
Tsengenes, Georgios A. Adamidis, B. S. K. K. Ibrahim, M. O. Tokhi, M. S. Huq, S. C. Gharooni,
Jos Luis Azcue, Alfeu J. Sguarezi Filho, Ernesto Ruppert

Published by InTech
Janeza Trdine 9, 51000 Rijeka, Croatia

Copyright 2012 InTech
All chapters are Open Access distributed under the Creative Commons Attribution 3.0 license,
which allows users to download, copy and build upon published articles even for commercial
purposes, as long as the author and publisher are properly credited, which ensures maximum
dissemination and a wider impact of our publications. After this work has been published by
InTech, authors have the right to republish it, in whole or part, in any publication of which they
are the author, and to make other personal use of the work. Any republication, referencing or
personal use of the work must explicitly identify the original source.

Notice
Statements and opinions expressed in the chapters are these of the individual contributors and
not necessarily those of the editors or publisher. No responsibility is accepted for the accuracy
of information contained in the published chapters. The publisher assumes no responsibility for
any damage or injury to persons or property arising out of the use of any materials,
instructions, methods or ideas contained in the book.

Publishing Process Manager Iva Simcic
Typesetting InTech Prepress, Novi Sad
Cover InTech Design Team

First published September, 2012
Printed in Croatia

A free online edition of this book is available at www.intechopen.com
Additional hard copies can be obtained from [email protected]

Fuzzy Controllers Recent Advances in Theory and Applications, Edited by Sohail Iqbal
p. cm.
ISBN 978-953-51-0759-0







Contents

Preface IX
Chapter 1 Fuzzy Logic Control of a Smart
Actuation System in a Morphing Wing 1
Teodor Lucian Grigorie, Ruxandra Mihaela Botez
and Andrei Vladimir Popov
Chapter 2 Embedded Fuzzy Logic Controllers
in Electric Railway Transportation Systems 23
Stela Rusu-Anghel and Lucian Gherman
Chapter 3 Design of a Real Coded GA Based Fuzzy Controller
for Speed Control of a Brushless DC Motor 63
Omer Aydogdu and Ramazan Akkaya
Chapter 4 A Type-2 Fuzzy Model Based
on Three Dimensional Membership Functions
for Smart Thresholding in Control Systems 85
M.H. Fazel Zarandi, Fereidoon Moghadas Nejad and H. Zakeri
Chapter 5 Fuzzy Control of Nonlinear Systems
with General Performance Criteria 119
Xin Wang, Edwin E. Yaz, James Long and Tim Miller
Chapter 6 A New Method for Tuning PID-Type Fuzzy Controllers
Using Particle Swarm Optimization 139
S. Bouallgue, J. Haggge and M. Benrejeb
Chapter 7 Output Tracking Control for Fuzzy Systems
via Static-Output Feedback Design 163
Meriem Nachidi and Ahmed El Hajjaji
Chapter 8 FPGA-Based Motion Control IC for Linear Motor
Drive X-Y Table Using Adaptive Fuzzy Control 181
Ying-Shieh Kung, Chung-Chun Huang and Liang-Chiao Huang
VI Contents

Chapter 9 Novel Yinger Learning Variable
Universe Fuzzy Controller 201
Ping Zhang and Guodong Gao
Chapter 10 Fuzzy Controllers: A Reliable Component
of Smart Sustainable Structural Systems 221
Maguid H. M. Hassan
Chapter 11 Vehicle Fault Tolerant Control Using
a Robust Output Fuzzy Controller Design 249
M. Chadli and A. El Hajjaji
Chapter 12 Wheelchair and Virtual Environment
Trainer by Intelligent Control 271
Pedro Ponce, Arturo Molina and Rafael Mendoza
Chapter 13 A Hybrid of Fuzzy and Fuzzy Self-Tuning PID Controller
for Servo Electro-Hydraulic System 299
Kwanchai Sinthipsomboon, Issaree Hunsacharoonroj,
Josept Khedari, Watcharin Po-ngaen and Pornjit Pratumsuwan
Chapter 14 Design and Simulation of Anfis Controller
for Virtual-Reality-Built Manipulator 315
Yousif I. Al Mashhadany
Chapter 15 Hierarchical Fuzzy Control 335
Carlos Andr Guerra Fonseca, Fbio Meneghetti Ugulino de Arajo
and Marconi Cmara Rodrigues
Chapter 16 Enhancing Fuzzy Controllers Using Generalized
Orthogonality Principle 367
Nora Boumella, Juan Carlos Figueroa and Sohail Iqbal
Chapter 17 New Areas in Fuzzy Application 385
Muhammad M.A.S. Mahmoud
Chapter 18 Fuzzy Control Systems: LMI-Based Design 441
Morteza Seidi, Marzieh Hajiaghamemar and Bruce Segee
Chapter 19 New Results on Robust

Filter
for Uncertain Fuzzy Descriptor Systems 465
Wudhichai Assawinchaichote
Chapter 20 Robust Stabilization for Uncertain
Takagi-Sugeno Fuzzy Continuous Model
with Time-Delay Based on Razumikhin Theorem 481
Yassine Manai and Mohamed Benrejeb
Contents VII

Chapter 21 A Two-Layered Load and Frequency
Controller of a Power System 503
Mavungu Masiala, Mohsen Ghribi and Azeddine Kaddouri
Chapter 22 Performance Evaluation of PI
and Fuzzy Controlled Power Electronic
Inverters for Power Quality Improvement 519
Georgios A. Tsengenes and Georgios A. Adamidis
Chapter 23 Discrete-Time Cycle-to-Cycle Fuzzy Logic
Control of FES-Induced Swinging Motion 541
B. S. K. K. Ibrahim, M. O. Tokhi, M. S. Huq and S. C. Gharooni
Chapter 24 Three Types of Fuzzy Controllers Applied
in High-Performance Electric Drives
for Three-Phase Induction Motors 559
Jos Luis Azcue, Alfeu J. Sguarezi Filho and Ernesto Ruppert







Preface

At the core of many engineering problems is the problem of control of different
systems. These systems range all the way from classical inverted pendulum to auto-
focusing system of a digital camera. Fuzzy control systems have demonstrated their
enhanced performance in all these areas. Although initially fuzzy systems were
associated only to the artificial intelligence that has refrained to the development of
theoretical fuzzy systems, in 1985 Japanese researchers Seiji Yasunobu and Soji
Miyamoto demonstrated the superiority of fuzzy control systems for the Sendai
railway. From that moment on, many applications have taken the advantage of the
inherent potential offered by fuzzy controllers. Some notable works on the
applications of fuzzy controllers are inverted pendulum balancing by Takeshi
Yamakawa, improved vacuum cleaners by Panasonic corporation, stable CCD
development by Canon Inc., energy efficient air conditioners by Mitsubishi Companies
and fuel efficient automatic space docking by NASA.
Since fuzzy controllers have proven their performance in many domains of science
and technology, it has led to further development of the theory of fuzzy systems to
solve even more intricate problems. In this book, our purpose is to present the recent
developments both in theory and applications of fuzzy controllers. The book is a
collection of chapters which are the result of the coordinated work of scholars
worldwide. Each chapter presents a different application of fuzzy controllers along
with the necessary development of the theory. Any reader can study every chapter of
this book as a self-contained research work. Moreover, this book can be recommended
to students who have done the basics of fuzzy set theory earlier on and now they want
to apply it. Reading of the entire book will provide you with a variety of ideas to
develop theory and apply it to fuzzy control problems.
This book starts with theory development and its application to solve an aerospace
engineering problem, then a fuzzy controller is used in electric railway transportation.
In the subsequent chapters, further theory and its applications to solve a variety of
problems are presented. These problems include the fault tolerant control of a vehicle,
integral wheelchair control, and hierarchical fuzzy control among others. Book
concludes with a chapter on describing the new areas in fuzzy controllers.
Reviewing all the received chapters, proposing improvements, and all the tasks of
book editing within few months were not possible without the dedicated efforts of my
X Preface

colleagues and co-editors of the book Nora Boumella and Juan Carlos Figueroa Garca.
I am thankful to both for their commitment to excellence.
During the review of the book chapters and book editing Iva Simcic, publishing
process manager of INTECH, provided fast and efficient feedback and guidance. I
would like to thank her for her professionalism.
Special thanks to my teachers Sarmad Abbasi and Yacine Amirat for guiding me to
understand the techniques of scientific research. I would also like to express my
gratitude to Arshad Ali, principal NUST School of Electrical Engineering and
Computer Science who always encourage faculty to take initiatives to promote a
culture of scientific investigation.
I am highly indebted to Higher Education Commission of Pakistan for revolutionizing
the science and technology. Moreover, I am also thankful to National University of
Sciences and Technology, Pakistan for providing an idea environment for research and
development.

Sohail Iqbal
National University of Science and Technology,
Pakistan

Chapter 1
Fuzzy Logic Control of a Smart
Actuation System in a Morphing Wing
Teodor Lucian Grigorie, Ruxandra Mihaela Botez and Andrei Vladimir Popov
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/48778
1. Introduction
The actual trends in aerospace engineering are related to the green aircrafts development
and to theirs' constructive parts optimization in order to obtain important fuel and energy
savings. A lot of these studies refer to the aircrafts' shape optimization, taking into
account that the aircraft drag force influences directly the fuel consumption. In this way, a
very interesting and provocative concept was launched on the market, i.e. "morphing
aircraft". Considering the drag reduction, fuel consumption economy and flight envelope
increasing promising benefits, many universities, R&D institutions and industry initiated
and developed morphing aircrafts studies in the last decade (Munday and Jacob, 2002;
Sanders, 2003; Manzo et al., 2004; Skillen and Crossley, 2005; Bornengo et al., 2005;
Moorhouse et al., 2006; Namgoong et al., 2006; Namgoong et al., 2007; Seigler et al., 2007;
Obradovic and Subbarao, 2011 a; Obradovic and Subbarao, 2011 b; Gamboa et al., 2009;
Baldelli at al., 2008; Inoyama et al., 2008; Thill et al., 2008; Perera and Guo, 2009; Bilgen et
al., 2009; Bilgen et al., 2010; Thill et al., 2010; Seber and Sakarya, 2010; Wildschek et al.,
2010; Ahmed et al., 2011). The multidisciplinary aspects involved by such studies, bring
together research teams in many fields of the science: aerodynamics and aeroelasticity,
automation, electrical engineering, materials engineering, control and software
engineering. Categorized as a part of the Smart structures engineering field, the general
concept of morphing aircrafts includes some particular elements, as a function by the
complexity of the developed morphing application. Recent researches in smart materials
and adaptive structures fields have led to a new way to obtain a morphing aircraft by
changing the shape of its wings through the control of the airfoils cambers; the concept
was called morphing wing. Therefore, a lot of architecture were and are still imagined,
designed, studied and developed, for this new concept application. One of these is our
team project including the numerical simulations and experimental multidisciplinary
studies using the wind tunnel for a morphing wing equipped with a flexible skin, smart

Fuzzy Controllers Recent Advances in Theory and Applications 2
material actuators and pressure sensors. The aim of these studies is to develop an
automatic system that, based on the information related to the pressure distribution along
the wing chord, moves the transition point from the laminar to the turbulent regime closer
to the trailing edge in order to obtain a larger laminar flow region, and, as a consequence,
a drag reduction.
The objective of the research presented here is to develop a new morphing mechanism using
smart materials such as Shape Memory Alloy (SMA) as actuators and fuzzy logic
techniques. These smart actuators deform the upper wing surface, made of a flexible skin, so
that the laminar-to-turbulent transition point moves closer to the wing trailing edge. The
ultimate goal of this research project is to achieve drag reduction as a function of flow
condition by changing the wing shape. The transition location detection is based on pressure
signals measured by optical and Kulite sensors installed on the upper wing flexible surface.
Depending on the project evolution phase, two architectures are considered for the
morphing system: open loop and closed loop. The difference between these two
architectures is their use of the transition point as a feedback signal. This research work was
a part of a morphing wing project developed by the Ecole de Technologie Suprieure in
Montral, Canada, in collaboration with the Ecole Polytechnique in Montral and the
Institute for Aerospace Research at the National Research Council Canada (IAR-NRC)
(Brailovski et al., 2008; Coutu et al., 2007; Coutu et al., 2009; Georges et al., 2009; Grigorie &
Botez, 2009; Grigorie & Botez, 2010; Grigorie et al., 2010 a; Grigorie et al., 2010 b; Grigorie et
al., 2010 c; Popov et al., 2008 a; Popov et al., 2008 b; Popov et al., 2009 a; Popov et al., 2009 b;
Popov et al., 2010 a; Popov et al., 2010 b; Popov et al., 2010 c; Sainmont et. al., 2009), initiated
and financially supported by the following government and industry associations: the
Consortium for Research and Innovation in Aerospace in Quebec (CRIAQ), the National
Sciences and Engineering Research Council of Canada (NSERC), Bombardier Aerospace,
Thales Avionics, and the National Research Council Canada Institute for Aerospace
Research (NRC-IAR).
2. Architecture of the controlled structure
To achieve the aerodynamic imposed purpose in the project, a first phase of the studies
involved the determination of some optimized airfoils available for 35 different flow
conditions (five Mach numbers and seven angles of attack combinations). The optimized
airfoils were derived from a laminar WTEA-TE1 reference airfoil (Khalid & Jones, 1993 a;
Khalid & Jones, 1993 b), and were used as a starting point for the actuation system
design.
The chosen wing model was a rectangular one, with a chord of 0.5 m and a span of 0.9 m.
The model was equipped with a flexible skin made of composite materials (layers of carbon
and Kevlar fibers in a resin matrix) morphed by two actuation lines (Fig. 1). Each of our
actuation lines uses three shape memory alloys wires (1.8 m in length) as actuators,
connected to a current controllable power supply. Also, each line contains a cam, which
moves in translation relative to the structure. The cam causes the movement of a rod related

Fuzzy Logic Control of a Smart Actuation System in a Morphing Wing 3
on the roller and on the skin. The recall used is a gas spring. So, when the SMA is heating
the actuator contracts and the cam moves to the right, resulting in the rise of the roller and
the displacement of the skin upwards. In contrast, the cooling of the SMA results in a
movement of the cam to the left, and thus a movement of the skin down. The horizontal
displacement of each actuator is converted into a vertical displacement at a rate 3:1 (results a
cam factor cf=1/3). From the optimized airfoils, an approximately 8 mm maximum vertical
displacement was obtained for the rods, so, a 24 mm maximum horizontal displacement
should be actuated.
In the same time, 32 pressure sensors (16 optical sensors and 16 Kulite sensors), were
disposed on the flexible skin in different positions along of the chord. The sensors are
positioned on two diagonal lines at an angle of 15 degrees from centreline (Fig. 2). The rigid
lower structure was made from Aluminium, and was designed to allow space for the
actuation system and wiring (Fig. 3).

Figure 1. Model of the flexible structure.
Starting from the reference airfoil, depending on different flow conditions, 35 optimized
airfoils were calculated for the desired morphed positions of the airfoil. The flow conditions
were established as combinations of seven incidence angles (-1, -0.5, 0, 0.5, 1, 1.5, 2) and
five Mach numbers (0.2, 0.225, 0.25, 0.275, 0.3). Each of the calculated optimized airfoils
should be able to keep the transition point as much as possible near the trailing edge.
airfoil
leading edge
Airfoil lower
surface (rigid)
Actuation
points
First
actuating line
From power
supply #1
From power
supply #2
Second
actuating line
Gas springs
SMA actuators
Airfoil trailing
edge
Airfoil uper surface - flexible
skin part (morphed)
Airfoil uper surface
(rigid part)
roller
rod
cam

Fuzzy Controllers Recent Advances in Theory and Applications 4

Figure 2. Pressure sensors distribution on the flexible skin

Figure 3. Cross section of the morphing wing model.
The SMA actuator wires are made of nickel-titanium, and contract like muscles when
electrically driven. Also, these have the ability to personalize the association of deflections
with the applied forces, providing in this way a variety of shapes and sizes extremely useful
to achieve actuation system goals. How the SMA wires provide high forces with the price of
small strains, to achieve the right balance between the forces and the deformations, required
by the actuation system, a compromise should be established. Therefore, the structural
components of the actuation system should be designed to respect the capabilities of
actuators to accommodate the required deflections and forces.
3. Open loop control of the morphing wing
For each of the two actuation lines the open loop control architecture used a controller
which took as a reference value the required displacement of the actuators from a database
stored in the computer memory to obtain the morphing wing optimized airfoil shape (Fig.
4); because the actuation lines structure was identical, both of them used the same
controller. As feedback signal the position signal from a linear variable differential
transducer (LVDT) connected to the oblique cam sliding rod of each actuator was used.
This method was called open-loop control due to the fact that this control method does
not take direct information from the pressure sensors concerning the wind flow
characteristics.
Chord wise
distribution of sensors
Actuator cam
SMA wires
Optical
sensors
Kulite
sensors
Leading edge
Trailing edge
Trailing
edge
Leading
edge
Cavities for instrumentation
Lower part
(rigid)
Actuators support
(rigid)
roller
Oblique cams
Uper surface
(flexible skin)
Actuation
point
rod
Actuation
lines
beam

Fuzzy Logic Control of a Smart Actuation System in a Morphing Wing 5

Figure 4. Open loop control architecture.
The SMA actuator control can be achieved using any method for position control. However,
the specific properties of SMA actuators such as hysteresis, the first cycle effect and the
impact of long-term changes must be considered.
Based on the 35 studied flight conditions, a database of the 35 optimized airfoils was built.
For each flight condition, a pair of optimal vertical deflections (dY1opt, dY2opt) for the two
actuation lines is apparent. The SMA actuators morphed the airfoil until the vertical
deflections of the two actuation lines (dY1real, dY2real) became equal to the required
deflections (dY1opt, dY2opt). The vertical deflections of the real airfoil at the actuation points
were measured using two position transducers. The controllers role is to send a command
to supply an electrical current signal to the SMA actuators, based on the error signals (e)
between the required vertical displacements and the obtained displacements. The designed
controller was valid for both actuation lines, which are practically identical.
From the point of view of the controller, the literature provides a lot of control techniques
for automatic systems. The global technology evolution has triggered an ever-increasing
complexity of applications, both in industry and in the scientific research fields. Many
researchers have concentrated their efforts on providing simple control algorithms to cope
with the increasing complexity of the controlled systems (Al-Odienat & Al-Lawama, 2008).
The main challenge of a control designer is to find a formal way to convert the knowledge
and experience of a system operator into a well-designed control algorithm (Kovacic &
Bogdan, 2006). From another point of view, a control design method should allow full
flexibility in the adjustment of the control surface, as the systems involved in practice are,
generally, complex, strongly nonlinear and often with poorly defined dynamics (Al-Odienat
& Al-Lawama, 2008). If a conventional control methodology, based on linear system theory,
is to be used, a linearized model of the nonlinear system should have been developed
beforehand. Because the validity of a linearized model is limited to a range around the
operating point, no guarantee of good performance can be provided by the obtained
controller. Therefore, to achieve satisfactory control of a complex nonlinear system, a
nonlinear controller should be developed (Al-Odienat & Al-Lawama, 2008; Hampel et al.,
2000; Kovacic & Bogdan, 2006; Verbruggen & Bruijn, 1997). From another perspective, if it
would be difficult to precisely describe the controlled system by conventional mathematical
Optimized
airfoils
database
Reference
airfoil
Flight condition
o, M (Re)
+
dY
1
dY
2
D
e
s
i
r
e
d

v
e
r
t
i
c
a
l
d
i
s
p
l
a
c
e
m
e
n
t
s
+
Controller
Power
Supply
SMA
actuator
Plant
LVDT
sensor
3:1 conversion
dX
1 measured
dY
1 measured
+
Controller
Power
Supply
SMA
actuator
Plant
LVDT
sensor
3:1 conversion
dX
2 measured
dY
2 measured
A
B
A B

Fuzzy Controllers Recent Advances in Theory and Applications 6
relations, the design of a controller using classical analytical methods would be totally
impractical (Hampel et al., 2000; Kovacic & Bogdan, 2006). Such systems have been the
motivation for developing a control system designed by a skilled operator, based on their
multi-year experience and knowledge of the static and dynamic characteristics of a system;
known as a Fuzzy Logic Controller (FLC) (Hampel et al., 2000). FLCs are based on fuzzy
logic theory, developed by L. Zadeh (Zadeh, 1965). By using multivalent fuzzy logic,
linguistic expressions in antecedent and consequent parts of IF-THEN rules describing the
operators actions can be efficiently converted into a fully-structured control algorithm
suitable for microcomputer implementation or implementation with specially-designed
fuzzy processors (Kovacic & Bogdan, 2006). In contrast to traditional linear and nonlinear
control theory, an FLC is not based on a mathematical model, and it does provide a certain
level of artificial intelligence compared to conventional PID controllers (Al-Odienat & Al-
Lawama, 2008).
Due to the strong non-linear character of the smart materials actuators used in our
application, one variant for the controller was developed by using the fuzzy logic
techniques. We tried to counterbalance the existence of a rigorous mathematical model, a
prior developed for system, avoiding in this way the loss of precision from linearization and
uncertainties in the systems parameters, which negatively influences the quality of the
resulting control. In the same time, we used the intuitive handling, simplicity and flexibility
capabilities offered by the fuzzy logic techniques and due to their closeness to human
perception and reasoning; fuzzy logic is an interface between logic an human reasoning,
providing an intuitive method for describing systems in human terms and automates the
conversion of those system specifications into effective models (Castellano et al., 2003;
Kovacic & Bogdan, 2006; Prasad Reddy et al., 2011; Zadeh, 1965).
The controller chosen structure was a PD fuzzy logic one, having as inputs the error
(difference between the desired and measured vertical displacement) and the change in
error (the derivative of the error), and as output the voltage controlling the Power Supply
output current (Fig. 5) (Kovacic & Bogdan, 2006). Widely accepted for capturing expert
knowledge, a Mamdani controller type was used, due to its simple structure of min-max
operations (Castellano et al., 2003).

Figure 5. Fuzzy controller architecture.
Fuzzy controller
Ae
e
u
Fuzzifier Defuzzifier Inference engine
Knowledge base

Fuzzy Logic Control of a Smart Actuation System in a Morphing Wing 7
The fuzzy controller internal mechanism during operation was relatively simple. On the
base of the membership functions (Fig. 6), stored in the knowledge base, the fuzzifier
converted the crisp inputs in linguistic variables. For our system, three membership
functions were chosen for both of the two inputs (N-negative, Z-zero, P-positive), while
five membership functions were considered for output (ZE-zero, PS-positive small, PM-
positive medium, PB-positive big, PVB-positive very big)(Fig. 6 and Table 1); the used
shape was the triangular one, defined by a lower limit a , an upper limit b , and a value m
( ) : a m b s s

0, if
, if
( )
, if
0, if
A
x a
x a
a x m
m a
x
b x
m x b
b m
x b

< s

< <

>

(1)
[-1, 1] interval was considered as universe of discourse for the two inputs, while for the
outputs was used [0, 1] interval.
Further, the inference engine converted the fuzzy inputs to the fuzzy output, based on the
If-Then type fuzzy rules in Table 2.
The fuzzified inputs were applied to the antecedents of the fuzzy rules by using the fuzzy
operator AND; in this way was obtained a single number, representing the result of the
antecedent evaluation. To obtain the output of each rule, the antecedent evaluation was
applied to the membership function of the consequent and the clipping (alpha-cut) method
was used; each consequent membership function was cut at the level of the antecedent truth.
Unifying the outputs of all eight rules, the aggregation process was performed and a fuzzy
set resulted for the output variable.

mf/
parameter
Input 1 (e) Input 2 (e) Output (u)
mf1 mf2 mf3 mf1 mf2 mf3 mf1 mf2 mf3 mf4 mf5
a -1 -1 0 -1 -0.5 0 0 0.1 0.3 0.6 0.8
m -1 0 1 -1 0 1 0 0.25 0.5 0.75 1
b 0 1 1 0 0.5 1 0.1 0.4 0.7 0.9 1
Table 1. Parameters of the input-output membership functions

e/e N Z P
N ZE ZE ZE
Z PS ZE ZE
P PM PVB PB
Table 2. Inference rules

Fuzzy Controllers Recent Advances in Theory and Applications 8

Figure 6. Membership functions
Because the output of the fuzzy system should be a crisp number, finally a defuzzification
process was realized (Fig. 7); the Centroid of area (COA) method was used. The control
surface resulted as in Fig. 8.
The fuzzy control surface was chosen in this way because it is normal that in the SMA
cooling phase the actuators would not be powered. Therefore, the fuzzy controller was
chosen to work in tandem with a bi-positional controller (particularly an on-off one). The
cooling phase may occur not only when controlling a long-term phase, when a switch
between two values of the actuator displacements is commended, but also in a short-lived
phase, which happens when the real value of the deformation exceeds its desired value and
the actuator wires need to be cooled. As a consequence, the final controller should behave as
a switch between the SMA cooling and heating phases, in which the output current is 0 A,
or is controlled by the fuzzy logic controller.
As a consequence, the resulted controller operational scheme can be organised as in Fig. 9.
To optimize all coefficients in the control scheme, the open loop of the morphing wing
system was implemented in Matlab-Simulink model as in Fig. 10.
D
e
g
r
e
e

o
f

m
e
m
b
e
r
s
h
i
p
0
0.2
0.4
0.6
0.8
1
-1 -0.5 0 0.5 1
input2 (Change in error)
mf1(ZE) mf2(PS) mf3(PM) mf4(PB) mf5(PVB)
D
e
g
r
e
e

o
f

m
e
m
b
e
r
s
h
i
p
0
0.2
0.4
0.6
0.8
1
output
0 0.2 0.4 0.6 0.8 1
input1 (Error)
D
e
g
r
e
e

o
f

m
e
m
b
e
r
s
h
i
p
mf1(N) mf2(Z) mf3(P)
0
0.2
0.4
0.6
0.8
1
-1 -0.5 0 0.5 1
mf1(N) mf2(Z) mf3(P)

Fuzzy Logic Control of a Smart Actuation System in a Morphing Wing 9

Figure 7. Fuzzy system operating mechanism.

Figure 8. Control surface.
The Mechanical system block implements all the forces influencing the SMA load force:
the aerodynamic force Faero, the skin force Fskin, and the gas spring force Fspring; in the
initialization phase, the actuators are preloaded by the gas springs even when there is no
aerodynamic load applied on the flexible skin.
The Fuzzy controller block models the controller presented in Fig. 9. Also, SMA actuators
physical limitations in terms of temperature and supplying currents were considered in this
block. Its detailed Simulink scheme is shown in Fig. 11. The block inputs are the control
-1
-0.5
0
0.5
1
-1
0
1
input1 (Error)
input2
(Change in error)
o
u
t
p
u
t
1
0.2
0.4
0.6
0.8

Fuzzy Controllers Recent Advances in Theory and Applications 10
error (the difference between the desired and the obtained displacements see Fig. 9) and
the SMA wires temperatures, while its output is the electrical current used to control the
actuators. The first switch assured the functioning in tandem of the fuzzy controller with the
on-off controller selecting one of the two options shown in Fig. 9 (error is positive or not),
while the second one protected the system by switching the electrical current value to 0A
when the SMA temperature value is over the imposed limit. As a supplementary protection
measure, a current saturation block was used to prevent the current from going over the
physical limit supported by the SMA wires.




Figure 9. Operational scheme of the controller.




Figure 10. Simulation model of the morphing wing system open loop.
K
P
Proportional
gain
e
K
D
Ae
Derivative
gain
K
O
Change in
output gain
e - Membership Functions
Ae - Membership Functions
Fuzzification Inference Defuzzification
s
N Z
N Z
Z PS
P PB
Z
-
Z
Ae / e
+
-
i
Command
e > 0
Desired Deflection
superior to
Actual Deflection ?
Error
i=0 for
heating phase
NO
YES
i=0 for
cooling phase
SMA
actuator
Real
deflection
e
Controller
skin deflection [mm]
desired deflection
[mm]
SMA elongation
[m]
1
1 7
desired
deflection
[mm]
3/1000
cam factor
mm to m
Scope
0.078
SMA max set
[m]
Current
Force
Displacement
Temperature
SMA Model
1.8
SMA Initial
length [m]
1.8
SMA Initial
length
Memory2
Memory1
F aero [N]
x [m]
F SMA [N]
y [mm]
Mechanical system
Diff error
Temperature
Current out
Fuzzy controller
273.15
Celsius to Kelvin
1150
Aerodynamic
force [N]

Fuzzy Logic Control of a Smart Actuation System in a Morphing Wing 11

Figure 11. Fuzzy controller block.
Another important block in the scheme in Fig. 10 is the SMA model block. This block
implemented a non-linear model for the SMA actuators using a Matlab S-function. The
model was built in the Shape Memory Alloys and Intelligent Systems Laboratory (LAMSI)
at ETS, using Lickhatchevs theoretical model (Terriault et al., 2006).
After a tuning operation the optimum values of the gains in the scheme were established.
Further, the controller was tested through numerical simulation to ensure that it works well.
Fig. 12 shows the response of the actuator relative to the desired vertical displacement, the
SMA actuator envelope (obtained vertical displacement vs. temperature), the SMA
temperature in time, and the SMA loading force vs. temperature. Using a preliminary
estimation of the forces loading the mechanical system, the next values were considered in
simulations: 1150 N for aerodynamic force; 1250 N for gas spring pretension force; and the
linear elastic coefficients of 2.95 N/mm and 100 N/mm, for the gas spring and for the flexible
skin, respectively.
The relative allure of the obtained and desired displacements, proved the good functioning
of the controller; the systems response is a critically damped one, an easier latency being
observed in the cooling phase of the SMA wires in comparison with theirs heating phase.
The SMAs temperature oscillations in the steady-state of the actuation position are due to
theirs thermal inertia, and do not affect significantly the SMA elongation. The shape of the
displacement vs. temperature and loading force vs. temperature envelopes highlights
the strong nonlinear behavior of the SMA actuators.
To validate the control some experimental tests in wind tunnel were performed; all tests
were performed in the IAR-NRC wind tunnel at Ottawa. The open loop experimental model
is presented in Fig. 13.
According to the architecture presented in Fig. 13, the controller acted on the SMA lines by
using a data acquisition card and two power supplies. The controller had also a feedback
from the SMA lines behavior by using the information from two position sensors. As power
supplies were chosen two Programmable Switching Power Supplies AMREL SPS100-33,
while a Quanser Q8 data acquisition card was used to interface them with the control
|u|
Abs
Current
1
Current out
Temperature
limiter
Switch
-K-
P Gain -K-
General
Gain
-1
Gain
Fuzzy Logic
Controller
du/dt -K-
D Gain
0
Current in
cooling phase
Current
saturation
0
Current
when reached
limit
2 Temperature
1
Diff error
Current
Mux

Fuzzy Controllers Recent Advances in Theory and Applications 12
software. The card was connected to a PC and programmed via Matlab/Simulink R2006b
and WinCon 5.2. The Matlab/Simulink implemented controller received the feedback signals
from two Linear Variable Differential Transformer (LVDT) potentiometers, used as position
sensors to monitor the SMA wires elongations. Also, as a safety feature for the experimental
model, the SMA wires temperatures were monitored and limited by the control system.
Therefore, as acquisition card inputs were considered the signals from the two LVDT
potentiometers and the six signals from the thermocouples installed on each of the SMA
wires components, while as outputs were considered 4 channels, used to initialize and to
control each power supply through theirs analog/external control features by means of a
DB-15 I/O connector.

Figure 12. Numerical simulation results
In the open loop wind tunnel tests, simultaneously with the controller validation, the real-
time detection and visualization of the transition point position were performed (Fig. 13),
for all the thirty-five optimized airfoils; a comparative study was realized based on the
transition point position estimation for the reference airfoil and for each optimized airfoil,
with the aim to validate the aerodynamic part of the project. In this way, the pressure data
signals obtained from the Kulite pressure sensors were used; these data were acquired using
-1
0
1
2
3
4
5
6
7
8
9
20 40 60 80 100 120
Temperature [
o
C]
V
e
r
t
i
c
a
l

d
i
s
p
l
a
c
e
m
e
n
t

d
Y

[
m
m
]
0 50 100 150 200 250 300
20
30
40
50
60
70
80
90
100
110
T
e
m
p
e
r
a
t
u
r
e

[
o
C
]
Time [s]
20 40 60 80 100 120
F
o
r
c
e

[
k
N
]
Temperature [
o
C]
0.85
0.9
0.95
1
1.05
1.1
1.15
1.2
1.25
0 50 100 150 200 250 300
-1
0
1
2
3
4
5
6
7
8
9
V
e
r
t
i
c
a
l

d
i
s
p
l
a
c
e
m
e
n
t

d
Y

[
m
m
]
Time [s]
desired
obtained
I
n
i
t
i
a
l
i
s
a
t
i
o
n

p
h
a
s
e
I
n
i
t
i
a
l
i
s
a
t
i
o
n

p
h
a
s
e

Fuzzy Logic Control of a Smart Actuation System in a Morphing Wing 13
the IAR-NRC analog data acquisition system, which was connected to the sensors. The
sampling rate of each channel was at 15 kHz, which allowed a pressure fluctuation FFT
spectral decomposition of up to 7.5 kHz for all channels. The signals were processed in real
time using Simulink. The pressure signals were analyzed using Fast Fourier Transforms
(FFT) decomposition to detect the magnitude of the noise in the surface air flow.
Subsequently, the data was filtered by means of high-pass filters and processed by
calculating the Root Mean Square (RMS) of the signal to obtain a plot diagram of the
pressure fluctuations in the flow boundary layer. This signal processing was necessary to
disparate the inherent electronically induced noise, by the Tollmien-Schlichting waves that
are responsible for triggering the transition from laminar to turbulent flow. The
measurements analysis revealed that the transition appeared at frequencies between 35kHz
and the magnitude of the pressure variations in the laminar flow boundary layer were on
the order of 5e-4 Pa. The transition from the laminar flow to turbulent flow was shown by
an increase in the pressure fluctuation, which was indicated by a drastic variation of the
pressure signal RMS.

Figure 13. Architecture of the open loop morphing wing model.
In Fig. 14 are presented the results obtained for the open loop controller testing in the flow
case characterized by M=0.275 and =1.5 deg (run test 51); can be easily observed that,
because of the gas springs pretension forces, the controller worked even the required
vertical displacements for the actuation lines were zero millimeters. Also, some noise
parasitizing the LVDT sensors measurements appeared in this test due to the wind tunnel
electrical power sources and its instrumentation equipment. The transition monitoring
revealed that this noise level did not influence significantly the transition point position; the
positioning resolution was determined by the density of the chord-disposed pressure
sensors.


Controller
Actuation
error
o, M,
Re
H
u
m
a
n
O
p
e
r
a
t
o
r
Flight
conditions
Optimised
airfoils
database
Computer
Desired
dY
1
, dY
2
(dY
1opt
, dY
2opt
)
dY
1
dY
2
From LVDT
position sensors
dY
1real
,
dY
2real
Temperatures
Reference airfoil
Real airfoil
(morphed)
From thermocouples
Electrical
current
Pressure
sensors
Power supplies
Gas
spring
Roller
SMA
Rod
Cam
LVDT
sensor
Transition position visualization
CH 1
CH 2
CH 3
CH 4
CH 5
CH 6
CH 7
CH 8
CH 9
CH 10
CH 13
CH 11
CH 12
Frequency [kHz]
M
a
g
n
i
t
u
d
e

[
d
B
]
Signal processing
Matlab/Simulink
Reference airfoil
Data acquisition system for pressure sensors
Thermocouples

Fuzzy Controllers Recent Advances in Theory and Applications 14


Figure 14. Wind tunnel test results for M=0.275 and =1.5 deg flow condition.
Fig. 15 depicts the results obtained by the transition monitoring for the run test 51 (M=0.275
and =1.5 deg); shown are the instant plots of the RMSs and spectrum for the pressure
signals channels with un-morphed and morphed airfoil.
From 16 Kulite pressure sensors initially mounted on the flexible skin, only 13 channels
were available (CH1 to CH13): sensor #1 was broken before the wind tunnel test, while the
sensors #12 and #13 were removed from plots due to the bad dynamic signals which show
electrical failure of the sensors. The left hand column presents the results for the reference
(un-morphed) airfoil, and the right hand side column display the results for the optimized
(morphed) airfoil. The spike of the RMS and the highest noise band on the spectral plots
(CH 11 cian spectra on the right low plot) for the morphed airfoil case suggested that the
flow was already turned turbulent on sensor on the channel 11 (eleventh available Kulite
sensor), near the trailing edge; therefore, the transition point position was somewhere near
the CH 11. For un-morphed airfoil the transition was localized by the sensor on the channel
8, with maximum RMS and the highest noise band on the spectral plots (CH 8 black spectra
on the left middle plot).
-1
0
1
2
3
4
5
6
7
8
V
e
r
t
i
c
a
l

d
i
s
p
l
a
c
e
m
e
n
t

d
Y
2

[
m
m
]
0 50 100 150 200 250 300 350
Time [s]
SMA2 obtained
SMA2 desired (7.40 mm)
SMA1 obtained
SMA1 desired (5.84 mm)
V
e
r
t
i
c
a
l

d
i
s
p
l
a
c
e
m
e
n
t

d
Y
1

[
m
m
]
-1
0
1
2
3
4
5
6
7
0 50 100 150 200 250 300 350
Time [s]
0 50 100 150 200 250 300 350
Time [s]
20
25
30
35
40
45
50
55
60
T
e
m
p
e
r
a
t
u
r
e

[
o
C
]
SMA2
SMA1
Temperature [
o
C]
V
e
r
t
i
c
a
l

d
i
s
p
l
a
c
e
m
e
n
t

[
m
m
]
SMA2
SMA1
-1
0
1
2
3
4
5
6
7
8
20 25 30 35 40 45 50 55 60

Fuzzy Logic Control of a Smart Actuation System in a Morphing Wing 15
The results obtained from the wind tunnel tests of open loop architecture showed that the
controller performed very well in enhancing the wind aerodynamic performance.

Figure 15. Transition monitoring in wind tunnel test for M=0.275 and =1.5 deg.
4. Closed loop control of the morphing wing
The next step of the work on the morphing wing project supposed the development of the
closed loop control, based on the pressure information received from the sensors and on the
transition point position estimation. The closed loop control included, as inner loop, the
actuation lines previous presented controller (Popov et al., 2010 a; Popov et al., 2010 b;
Popov et al., 2010 c).
The closed loop architecture was developed in order to generate real time optimized airfoils
starting from the information received from the pressure sensors and targeting the
morphing wing main goal: the improvement of the laminar flow over the wing upper
surface (Fig. 16); the previously calculated optimized airfoils database was by-passed in this
control strategy, and were used just to see if the closed loop real time optimizer conducted
to similar results for morphed airfoil in a flow case. To achieve the control, a mixed
optimization method was used, between the gradient ascent or hill climbing method
Morphed airfoil
Channels 1 to 5
Channels 6 to 10
Channels 11 to 13
Channels 1 to 5
Channels 6 to 10
Channels 11 to 13
CH 1 CH 2 CH 3
CH 4
CH 5
CH 6
CH 7
CH 8
CH 9
CH 10
CH 12
CH 11
CH 13
disabled sensors broken
sensor
1
0.8
0.6
0.4
0.2
0
R
M
S
Sensor #
2 4 6 9 11 13 15 1 3 5 7 8 10 12 14 16
13 available pressure sensors
R
M
S
CH 1 CH 2 CH 3
CH 4
CH 5
CH 6
CH 7
CH 8
CH 9
CH 10
CH 12
CH 11
CH 13
broken sensor
disabled sensors
1
0.8
0.6
0.4
0.2
0
Sensor #
2 4 6 9 11 13 15 1 3 5 7 8 10 12 14 16
Eleventh sensor
Eleventh sensor
13 available pressure sensors
Turbulence
Unmorphed airfoil

Fuzzy Controllers Recent Advances in Theory and Applications 16
and the simulated annealing method. Two variants were tested for the starting point on
the optimization map control: 1) dY1=4 mm, dY2=4 mm (Fig. 16), and 2) dY1opt, dY2opt of the
theoretically obtained optimized airfoil (Popov et al., 2010 a; Popov et al., 2010 b; Popov et
al., 2010 c).

Figure 16. Optimization logic scheme for closed loop.
For the new control architecture, the software application was developed in
Matlab/Simulink and two National Instruments Data Acquisition Cards were used: NI-DAQ
USB 6210 and NI-DAQ USB 6229 (Quanser Q8 data acquisition card was removed from this
configuration). As feedback signal for control was used the transition point position
estimated starting from the pressure signals from the Kulite sensors. In the beginning of
wind-tunnel tests, a number of sixteen Kulite sensors were installed, but due to their
removal and re-installation during the next two wind tunnel tests, four of them were found
defective. Therefore, a number of twelve sensors remained to be used during the last wind
tunnel tests.
The closed loop control results and the followed optimization trajectory for =0.5 and
M=0.3 flow case are shown in Fig. 17. In this case, as starting point in optimization was used
the point with the coordinates dY1opt and dY2opt, characterizing the theoretically obtained
optimized airfoil: dY1opt=4.81 mm, and dY2opt=7.45 mm. The obtained rezults validated the
theoretical optimized airfoil obtained by Ecole Polytechnique in Montreal for this flow case,
taking into account that optimization method implemented in the closed loop conducted to
a morphed airfoil almost identical with the first one (dY1opt_cl=4.66 mm, and dY2opt_cl =7.28
mm), and the transition was detected on the same pressure sensor with the open loop case
(the tenth sensor in the array).
In Figs 18 and 19 are presented the FFTs of the Kulite pressure sensors data, and the
pressure data RMSs for un-morphed (reference) and closed loop real time optimized airfoils,
in this flow case. In Fig. 19 can be also observed the N factor (for transition positioning)
distribution for reference airfoil and optimized airfoil. The distribution was estimated by
using the XFoil computational fluid dynamics; XFoil code is free licensed software in which
the e
n
transition criterion is used (Drela, 2003; Drela and Giles, 1987). In these graphs, the N
values calculated by XFoil for various sensors are defined by circles. In the morphed-to-
optimized airfoil case, the RMS plot displayed in Fig. 19 with star symbols, showed that the
sensor with the maximum RMS has become the tenth sensor plotted
Start optimization
dY
1
=4, dY
2
=4
Optimizer
Find max(x
tr
)
B-spline model
+
WTEA airfoil
Flow conditions
o, M (Re)
dY
1
dY
2
New
morphed
airfoil
XFoil
CFD code
x
tr
Transition
position
C
p
Distribution

Fuzzy Logic Control of a Smart Actuation System in a Morphing Wing 17

Figure 17. The closed loop real time optimization results for =0.5 and M=0.3 flow case.
The spectral decomposition of the pressure signals in Fig. 18 confirmed the Tollmien
Schlichting waves occurrence in the tenth sensor, visible in the highest power spectra
(twelfth channel in the right hand side plots) in the frequency band of 25 kHz.

Figure 18. Pressure signals FFT for un-morphed and real time optimized airfoils, for = 0.5 and M=0.3.

Figure 19. The pressure data RMSs and the N factor distribution
6
7
8
3
4
5
6
7
P1
Optimization trajectory
dY
2
[mm]
dY
1
[mm]
0
2
4
6
8
10
transition on sensor 8
transition on sensor 9
transition on sensor 10 transition on sensor 6
optimum position
Turbulence
Tenth sensor
Unmorphed airfoil Morphed airfoil
Turbulence
Tenth
sensor
Available pressure sensors
RMS curve
RMS curve
Unmorphed airfoil Real time optimized airfoil

Fuzzy Controllers Recent Advances in Theory and Applications 18
5. Conclusions
The design and validation results for an actuation system of a morphing wing were
exposed. The developed morphing mechanism used smart materials such as Shape
Memory Alloy (SMA) in the actuation mechanism. Two architectures were developed for
the used control system: an open loop, and a closed loop one. The open loop architecture of
the controller was used as an inner loop of the closed loop structure, and included a PD
fuzzy logic controler in tandem with an on-off clasical controller. Both of the control
architectures were validated in wind tunnel tests in parallel with the transition point real
time position detection and visualization. In the closed loop controller architecture, the
information about the external airflow state received from the pressure sensors system was
considered and the decisions have been taken based on the transition point position
estimation.
Author details
Teodor Lucian Grigorie, Ruxandra Mihaela Botez and Andrei Vladimir Popov
cole de Technologie Suprieure, Canada
Acknowledgement
We would like to thank the Consortium of Research in the Aerospatial Industry in Quebec
(CRIAQ), Thales Avionics, Bombardier Aerospace, and the National Sciences and
Engineering Research Council (NSERC) for the support that made this research possible. We
would also like to thank George Henri Simon for initiating the CRIAQ 7.1 project, and
Philippe Molaret from Thales Avionics and Eric Laurendeau from Bombardier Aeronautics
for their collaboration on this work.
6. References
Ahmed, M.R.; Abdelrahman, M.M.; ElBayoumi, G. M. & ElNomrossy, M.M. (2011). Optimal
wing twist distribution for roll control of MAVs, The Aeronautical Journal, Royal
Aeronautical Society, vol. 115, 2011, pp. 641-649, ISSN: 0001-9240
Al-Odienat, A.I. & Al-Lawama, A.A. (2008). The Advantages of PID Fuzzy Controllers Over
The Conventional Types, American Journal of Applied Sciences, Vol. 5, No. 6, pp. 653-658,
June 2008, ISSN: 1546-9239
Baldelli, D.H.; Lee, D.H.; Sanchez Pena R.S. & Cannon, B. (2008). Modeling and Control of
an Aeroelastic Morphing Vehicle, Journal of Guidance, Control, and Dynamics, Vol. 31, No.
6, NovemberDecember 2008, pp. 1687-1699, ISSN: 0731-5090
Bilgen, O.; Kochersberger, K.B. & Inman, D.J. (2009). Macro-Fiber Composite Actuators for a
Swept Wing Unmanned Aircraft, The Aeronautical Journal, Royal Aeronautical Society,
Vol. 113, 2009, pp. 385-395, ISSN: 0001-9240

Fuzzy Logic Control of a Smart Actuation System in a Morphing Wing 19
Bilgen, O.; Kochersberger, K.B.; Inman, D.J. & Ohanian, O.J. (2010). Novel, Bidirectional,
Variable-Camber Airfoil via Macro-Fiber Composite Actuators, Journal of Aircraft, Vol.
47, No. 1, JanuaryFebruary 2010, pp. 303-314, ISSN: 0021-8669
Bornengo, D.; Scarpa, F. & Remillat, C. (2005). Evaluation of hexagonal chiral structure for
morphing airfoil concept, Proceedings of the Institution of Mechanical Engineers, Part G:
Journal of Aerospace Engineering, 2005, vol. 219, 3, pp. 185-192, ISSN: 0954-4100
Brailovski, V.; Terriault, P.; Coutu, D.; Georges, T.; Morellon, E.; Fischer, C. & Berube, S.
(2008). Morphing laminar wing with flexible extrados powered by shape memory alloy
actuators, Proceedings of ASME 2008 Conference on Smart Materials, Adaptive
Structures and Intelligent Systems (SMASIS2008), pp. 615-623, ISBN: 978-0-7918-4331-4,
Maryland, USA, October 2830, 2008, Publisher ASME, Ellicott City
Castellano, G.; Fanelli, A. M. & Mencar, C. (2003). Design of transparent mamdani fuzzy
inference systems. In Design and application of hybrid intelligent systems book, IOS Print,
Amsterdam, pp. 468476, ISBN:1-58603-394-8
Coutu, D.; Brailovski, V.; Terriault, P. & Fischer, C. (2007). Experimental validation of the 3D
numerical model for an adaptive laminar wing with flexible extrados, Proceedings of
the 18th International Conference of Adaptive Structures and Technologies, 10 pages,
Ottawa, Ontario, Canada, 3-5 October, 2007
Coutu, D.; Brailovski, V. & Terriault, P. (2009). Promising benefits of an active-extrados
morphing laminar wing, AIAA Journal of Aircraft, Vol. 46, No. 2, pp. 730-731, March-
April 2009, ISSN: 0021-8669
Drela, M. (2003). Implicit Implementation of the Full en Transition Criterion, 21st Applied
Aerodynamics Conference, Orlando, Florida, 2326 June 2003, pp. 18.
Drela, M. & Giles, M.B. (1987). Viscous-Inviscid Analysis of Transonic and Low Reynolds
Number Airfoils, Journal of Aircraft, Vol. 25, No. 10, 1987, pp. 13471355, ISSN: 0021-
8669
Gamboa, P.; Vale, J.; Lau, F. J. P. & Suleman, A. (2009). Optimization of a Morphing Wing
Based on Coupled Aerodynamic and Structural Constraints, AIAA Journal, Vol. 47, No.
9, September 2009, pp. 2087-2104, ISSN: 0001-1452
Georges, T.; Brailovski, V.; Morellon, E.; Coutu, D. & Terriault, P. (2009). Design of Shape
Memory Alloy Actuators for Morphing Laminar Wing With Flexible Extrados, Journal
of Mechanical Design, Vol. 131, No. 9, 9 pages, 091006, September 2009, ISSN: 1050-
0472
Grigorie, T.L. & Botez, R.M. (2009). Adaptive neuro-fuzzy inference system based
controllers for Smart Material Actuator modeling, Proceedings of the Institution of
Mechanical Engineers, Part G: Journal of Aerospace Engineering, Vol. 223, No. 6, pp. 655-
668, June 2009, ISSN: 0954-4100
Grigorie, T.L. & Botez, R.M. (2010). New adaptive controller method for SMA hysteresis
modeling of a morphing wing, The Aeronautical Journal, Vol. 114, No. 1151, pp. 1-13,
January 2010, ISSN: 0001-9240

Fuzzy Controllers Recent Advances in Theory and Applications 20
Grigorie, T.L.; Popov, A.V.; Botez, R.M.; Mbarki, Y. & Mamou, M. (2010 a). Modeling and
testing of a morphing wing in open-loop architecture, AIAA Journal of Aircraft, Vol. 47,
No. 3, pp. 917-923, MayJune 2010, ISSN: 0021-8669
Grigorie, T. L.; Popov, A.V.; Botez, R.M.; Mamou, M. & Mebarki, Y. (2010 b). A morphing
wing used shape memory alloy actuators new control technique with bi-positional and
PI laws optimum combination. Part 1: design phase, Proceedings of the 7th
International Conference on Informatics in Control, Automation and Robotics ICINCO
2010, pp. 5-12, ISBN: 978-989-8425-00-3, Madeira, Portugal, 15-18 June, 2010, SciTePress
Science and Technology Publications, Funchal
Grigorie, T. L.; Popov, A.V.; Botez, R.M.; Mamou, M. & Mebarki, Y. (2010 c). A morphing
wing used shape memory alloy actuators new control technique with bi-positional and
PI laws optimum combination. Part 2: experimental validation, Proceedings of the 7th
International Conference on Informatics in Control, Automation and Robotics ICINCO
2010, pp. 13-19, ISBN: 978-989-8425-00-3, Madeira, Portugal, 15-18 June, 2010,
SciTePress Science and Technology Publications, Funchal
Hampel, R.; Wagenknecht, M. & Chaker, N. (2000). Fuzzy Control Theory and Practice,
Physica-Verlag, ISBN-13: 978-3790813272, USA
Inoyama, D.; Sanders, B.P. & Joo, J.J. (2008). Topology Optimization Approach for the
Determination of the Multiple-Configuration Morphing Wing Structure, Journal of
Aircraft, Vol. 45, No. 6, NovemberDecember 2008, pp. 1853-1863, ISSN: 0021-8669
Khalid, M. & Jones, D.J. (1993). Navier Stokes Investigation of Blunt Trailing Edge Airfoils
using O-Grids, AIAA Journal of Aircraft, vol.30, no.5, pp. 797-800, 1993, ISSN: 0021-8669
Khalid, M. & Jones, D.J. (1993). A CFD Investigation of the Blunt Trailing Edge Airfoils in
Transonic Flow, Inaugural Conference of the CFD Society of Canada
Kovacic, Z. & Bogdan, S. (2006). Fuzzy Controller Design Theory and applications, Taylor and
Francis Group, ISBN: 978-0849337475, USA
Manzo, J.; Garcia, E. & Wickenheiser, A.M. (2004) Adaptive Structural Systems and
Compliant Skin Technology of Morphing Aircraft Structures, Proceedings of SPIE: The
International Society for Optical Engineering, Vol. 5390, 2004, pp. 225234
Moorhouse D.J.; Sanders B.; von Spakovsky, M.R. & Butt, J. (2006). Benefits and Design
Challenges of Adaptive Structures for Morphing Aircraft, The Aeronautical Journal,
Royal Aeronautical Society, vol. 110, 2006, pp. 157-162, ISSN: 0001-9240
Munday, D. & Jacob, J.D. (2002). Active Control of Separation on a Wing with Conformal
Camber, AIAA Journal of Aircraft, 39, No. 1, ISSN: 0021-8669
Namgoong, H.; Crossley, W.A. & Lyrintzis, A.S. (2006). Morphing Airfoil Design for
Minimum Aerodynamic Drag and Actuation Energy Including Aerodynamic Work,
AIAA Paper 2006-2041, 2006, pp. 54075421
Namgoong, H.; Crossley, W.A. & and Lyrintzis, A.S. (2007). Aerodynamic Optimization of a
Morphing Airfoil Using Energy as an Objective, AIAA Journal, Vol. 45, No. 9, September
2007, pp. 2113-2124, ISSN: 0001-1452

Fuzzy Logic Control of a Smart Actuation System in a Morphing Wing 21
Obradovic, B. & Subbarao, K. (2011 a). Modeling of Dynamic Loading of Morphing-Wing
Aircraft, Journal of Aircraft, Vol. 48, No. 2, MarchApril 2011, pp. 424-435, ISSN: 0021-
8669
Obradovic, B. & Subbarao, K. (2011 b). Modeling of Flight Dynamics of Morphing-Wing
Aircraft, Journal of Aircraft, Vol. 48, No. 2, MarchApril 2011, pp. 391-402, ISSN: 0021-
8669
Perera, M. & Guo, S. (2009). Optimal design of an aeroelastic wing structure with seamless
control surfaces, Proceedings of the Institution of Mechanical Engineers, Part G: Journal of
Aerospace Engineering, August 1, 2009, vol. 223, 8, pp. 1141-1151, ISSN: 0954-4100
Popov, A.V.; Botez, R.M. & Labib, M. (2008 a). Transition point detection from the surface
pressure distribution for controller design, AIAA Journal of Aircraft, Vol. 45, No. 1, pp.
23-28, January-February 2008, ISSN: 0021-8669
Popov, A.V.; Labib, M.; Fays, J. & Botez, R.M. (2008 b). Closed loop control simulations on a
morphing laminar airfoil using shape memory alloys actuators, AIAA Journal of Aircraft,
Vol. 45, No. 5, pp. 1794-1803, September-October 2008, ISSN: 0021-8669
Popov, A.V.; Botez, R.M.; Mamou, M.; Mebarki, Y.; Jahrhaus, B.; Khalid. M. & Grigorie, T.L.
(2009 a). Drag reduction by improving laminar flows past morphing configurations,
AVT-168 NATO Symposium on the Morphing Vehicles, 12 pages, 20-23 April, 2009,
Published by NATO, Evora, Portugal
Popov, A.V.; Botez, R. M.; Mamou, M. & Grigorie, T.L. (2009 b). Optical sensor pressure
measurements variations with temperature in wind tunnel testing, AIAA Journal of
Aircraft, Vol. 46, No. 4, pp. 1314-1318, July-August 2009, ISSN: 0021-8669
Popov, A.V.; Grigorie, T. L.; Botez, R.M.; Mamou, M. & Mebarki, Y. (2010 a). Morphing wing
real time optimization in wind tunnel tests, Proceedings of the 7th International
Conference on Informatics in Control, Automation and Robotics ICINCO 2010, pp. 114-
124, ISBN: 978-989-8425-00-3, Madeira, Portugal, 15-18 June, 2010, SciTePress Science
and Technology Publications, Funchal
Popov, A.V.; Grigorie, T. L.; Botez, R.M.; Mamou, M. & Mebarki, Y. (2010 b). Closed-Loop
Control Validation of a Morphing Wing Using Wind Tunnel Tests, AIAA Journal of
Aircraft, Vol. 47, No. 4, pp. 1309-1317, JulyAugust 2010, ISSN: 0021-8669
Popov, A.V.; Grigorie, T. L.; Botez, R.M.; Mamou, M. & Mebarki, Y. (2010 c). Real Time
Morphing Wing Optimization Validation Using Wind-Tunnel Tests, AIAA Journal of
Aircraft, Vol. 47, No. 4, pp. 1346-1355, JulyAugust 2010, ISSN: 0021-8669
Prasad Reddy, P.V.G.D. & Hari, Gh.V.M.K. (2011). Fuzzy Based PSO for Software Effort
Estimation, International Conference on Information Technology and Mobile
Communication (AIM 2011), Nagpur, India, Volume 147, Part 2, April 2011, pp. 227-232,
ISSN: 1865-0929
Sainmont, C.; Paraschivoiu, I. & Coutu, D. (2009). Multidisciplinary Approach for the
Optimization of a Laminar Airfoil Equipped with a Morphing Upper Surface, AVT-168
NATO Symposium on the Morphing Vehicles, 20-23 April, 2009, Published by NATO,
Evora, Portugal

Fuzzy Controllers Recent Advances in Theory and Applications 22
Sanders, B. (2003). Aerodynamic and Aeroelastic Characteristics of Wings with Conformal
Control Surfaces for Morphing Aircraft, Journal of Aircraft, Vol. 40, No. 1, 2003, pp. 94
99, ISSN: 0021-8669
Seber, G. & Sakarya, E. (2010). Nonlinear Modeling and Aeroelastic Analysis of an Adaptive
Camber Wing, AIAA Journal of Aircraft, Vol. 47, No. 6, NovemberDecember 2010, pp.
2067-2074, ISSN: 0021-8669
Seigler, T.M.; Neal, D.A.; Bae, J.S. & Inman, D.J. (2007). Modeling and Flight Control of
Large-Scale Morphing Aircraft, Journal of Aircraft, Vol. 44, No. 4, JulyAugust 2007, pp.
1077-1087, ISSN: 0021-8669
Skillen, M.D. & Crossley, W.A. (2005). Developing Response Surface Based Wing Weight
Equations for Conceptual Morphing Aircraft Sizing, AIAA Paper 2005-1960, 2005, pp.
20072019
Terriault, P.; Viens, F. & Brailovski, V. (2006). Non-isothermal Finite Element Modeling of a
Shape Memory Alloy Actuator Using ANSYS, Computational Materials Science, Vol. 36,
No. 4, July 2006, pp. 397-410, ISSN: 0927-0256
Thill, C.; Etches, J.; Bond, I.P.; Potter, K.D. & Weaver, P.M. (2008). Morphing skins review,
The Aeronautical Journal, Royal Aeronautical Society, Vol. 112, 2008, pp. 117-139, ISSN:
0001-9240
Thill, C.; Downsborough, J.D.; Lai, J.S.; Bond, I.P. & Jones, D.P. (2010). Aerodynamic study
of corrugated skins for morphing wing applications, The Aeronautical Journal, Royal
Aeronautical Society, vol. 114, 2010, pp. 237-244, ISSN: 0001-9240
Verbruggen, H.B. & Bruijn, P.M. (1997). Fuzzy control and conventional control: What is
(and can be) the real contribution of Fuzzy Systems?, Fuzzy Sets Systems, Vol. 90, No. 2,
pp. 151160, September 1997, ISSN: 0165-0114
Wildschek, A.; Havar, T. & Pltner, K. (2010). An all-composite, all-electric, morphing
trailing edge device for flight control on a blended-wing-body airliner, Proceedings of
the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering, January
2010, vol. 224, 1, pp. 1-9, ISSN: 0954-4100
Zadeh, L.A. (1965). Fuzzy sets, Information and Control, Vol. 8, No. 3, pp. 339-353, June 1965,
ISSN: 00199958
Chapter 2
Embedded Fuzzy Logic Controllers in Electric
Railway Transportation Systems
Stela Rusu-Anghel and Lucian Gherman
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/48588
1. Introduction
A. Power system harmonic pollution limitation using Fuzzy Logic controlled
active filters
1.1. Introduction
Nonlinear loads system has been grooving on influence in electric power due to the
advance of power electronics technologies. As a result, the harmonic pollution in the
power system deteriorates the power quality significantly. One effect of the harmonic
pollution is the harmonic resonance which may result in major voltage distortion in the
power system.
1.1.1. Harmonic effect
The current harmonic components could cause the following problems:
- resonance effect with overvoltage and overcurrent consequences,
- additional losses,
- psophomentic disturbance of the telecommunication systems,
- disturbance in the remote control systems,
- malfunction of protection devices,
- misoperation of semiconductor-controllers.
The harmonic disturbance basically could be characterized by the individual (1) and total (2)
harmonic distortion factors:

1
k
k
X
D
X
= (1)

Fuzzy Controllers Recent Advances in Theory and Applications 24

2
2
1
k
k
X
X
THD
X

=
=

(2)
where:
50
f
k
Hz
= : the harmonic order;
Xk : k
th
harmonic component of I or V;
Xi : fundamental frequency component of I or V
1.1.2. Psophometric interference
The high power lines could influence the neighbouring telecommunication networks by the
following ways:
- Capacitive coupling: The voltage of the power line causes charging current ;
- Inductive coupling: The line current induces longitudinal emf.
The most dominant part of the psophometric noise is the inducing effect caused by the zero
sequence components of the current. The power balance of the three-phase is near
symmetrical during normal operation, thus the coupling is measurable only if the distance
between the two systems is comparable with the phase distance of one system. However
electric traction is a single-phase system with ground return and in consequence it is a natural
zero sequence system. That is why it is important to calculate the psophometric noise. [1]
By telecommunication lines the rate of the disturbance could be characterized by the so
called psophometric voltage. It could be calculated by this formula:

2
800
f
p f
f
p
V V
p
| |
| =
|
\ .

(3)
where: Vf : voltage component by f frequency; Pf : psophometric weigh by f frequency;
P800 = 1000.
The psophometric weight has been determined after human tests; it could be seen on Fig. 1.
It could be concluded that the main part of the noise disturbance is caused by the 800 Hz
and surrounding harmonics. The psophometric weighting could be applied for the current
components, too, the formula is the same like in (3), however, this value is characteristic to
the zero sequence current of power line regarding its possible disturbing effect. This is the
so called disturbing current [1].
1.1.3. Active filtering
Several researchers propose the installation of active filter in order to damp the harmonic
resonance effect. The magnitude of damping provided by the active filter, the level of

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 25
harmonic distortion, may become worse in certain locations along the radial line. One
solution is to use multiple active filters located in the proximity of nonlinear load element.
In case of railway transportation, the power system pollution is mainly originated by the use
of DC locomotives equipped with rectifier units (fig. 2).

Figure 1. The psophometric weight
The harmonic filters are mandatory to limit the harmonic currents flowing into upstream
network and to decrease the resonance effect causing current amplification along the 25 kV
supply line. The combination of power factor correction capacitors, parasitic capacitance of
contact line and the system inductance (power cables, transformers, etc.) often result in
resonant frequency in the 600 800 Hz range.

Figure 2. Electrical diagram of the EA 060 locomotive used in Romania
Most active filter technologies, which focus on compensating harmonic current of nonlinear
loads, can not adequately address this issue.

Fuzzy Controllers Recent Advances in Theory and Applications 26
We propose the application of active filters in order to limit the harmonic currents produced
by the traction system. The active filter could be located in locomotive or on the substation,
or both. Coordination of harmonic of multiple filter units may become a problem since the
railway transportation system is characterized by the presence of different types of
locomotive from different ages of technology (DC motor with rectifier unit, thyristor).
However, in differently from the type of locomotive, the harmonic production needs to be
eliminated or limited to a acceptable value imposed by international standards.
In [2] a solution for the coordination of multiple active filters is proposed. The active filter
units which are placed on different locations can perform the harmonic filtering without a
direct communication using the droop characteristic. We implement the same solution using
a fuzzy logic control system.
1.2. Power quality in railway transportation
Power distribution system in electric railway transportation is presented in fig. 3.

Figure 3. Power distribution system

Figure 4. ST load current and voltage waveform in power substation

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 27

Figure 5. Harmonic spectrum of load current in power substation

Figure 6. Harmonic spectrum of load current in locomotive
In fig. 4, we present the waveforms for a work regime recorded in a real substation. In fig.s 5
and 6, is presented the harmonic spectrum for the load current in power substation and in
locomotive. Comparing these two different spectrums, it could be concluded that the
resonance effect is the highest at the 15
th
and 17
th
harmonics. Over the 25
th
harmonics the
supply system is decreasing the harmonic current. THD becomes to 34%, far exceeding the
admissible values. The resonance phenomenon increase psophometric interference.
1.3. Active filtering solutions for railway power systems
For harmonic compensations in case of railway applications the best choice is the single
phase bridge inverter with PWM controlled current control. In order to perform the
harmonic compensations it is necessary to present the control structure of the active power
filter. The control method is based on instantaneous power theory [3], [4] and [5].
The single phase power system can be defined using:
( ) cos( ) ( ) cos( ) u t U t i t I t e e = = (4)
In order to perform the orthogonal transformation of the single phase system to a
synchronous reference frame a fictitious imaginary phase defined as is introduced:
( ) sin( ) ( ) sin( )
i i i i
u t U t i t I t e e = = (5)
we obtain an orthogonal coordinate system with:

Fuzzy Controllers Recent Advances in Theory and Applications 28
( ) and ( )
i
u u t u u t
o |
= = (6)
The active power PAV and PiAV is given by:

4
0
2
2
T
AV
AV
P
P u i u i dt
T
o|
o o | |
| |
(
= = +
|

\ .
}
(7)

4
0
2
2
AV
AV
T
i
i
P
P u i u i dt
T
o|
o o | |
| |
(
= = +
|

\ .
}
(8)
The instantaneous power is:

( )
( )
p u i u i a
q u i u i b
o| o o | |
o| o o | |
= +
=
(9)
And the power factor is given by:

q
arctg
p
o|
o|

| |
| =
|
\ .
(10)
Using p-q-r power theory introduced by Kim and Akagi, allows to present the power
situation in synchronous rotation frame. In case we have:

2 2
u u u
o| o |
= + (11)
and

1
tan
u
u
|
o
u

| |
= |
|
\ .
(12)

0
1
0 0 0 0
0 0 0
0
p
q
r
u u u u
u
u u
u
u
u
o | o|
o
|
o|
o|
( ( ( (
( ( ( (
= =
( ( ( (
( ( ( (
(
(13)
and similarly

0
1
0 0 0 0
0 0 0
0
p
q
r
i u u i
i
i i
u
u
i
o | o|
o
|
o|
o|
( ( ( (
( ( ( (
= =
( ( ( (
( ( ( (
(
(14)
where:

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 29

p
i
u
o|
o|
o|
=
(15)
where as:

p u i u i p
o| o o o o
= + =
(16)
Finally we can state:

( )
( )
0 ( )
( )
p
p
q r q r
p p
u u a
i i b
u u i i c
p u i d
o|
o|
=
=
= = =
=
(17)
From (9a,b) can be obtain

~
2
~
1
AV
AV
u u i P p
i Q q u u
u
o | o
| | o
o|
(
( +
= (
(
+
(


(18)
In order to realize the compensation of whole reactive and distortion powers is necessary to
compensate QAV component as well as to filter q~ and p~ components. The reference
current is:

( ) ( )
~
2
1
REF AV AV
i u p P u Q q
u
o |
o|
(
= +

(19)
In fig. 7 we present the active power filter operation considered in two cases: substation
placed AF and locomotive placed AF.

Figure 7. Substation current without filtering at 25 kV and the current in 110 kV network
As it can be observed the placement of active filters in only one place can not provide the
necessary filtering in the power line feeding the substation simultaneously to the contact
line filtering.
Due to this fact it is necessary to adopt different control strategy and solutions.

Fuzzy Controllers Recent Advances in Theory and Applications 30

Figure 8. Substation Harmonics without filtering

Figure 9. Substation AFU waveforms in substation and in locomotive unit

Figure 10. Harmonic components for active filter placed
1.3.1. Coordination of multiple active filters in railway power systems
We propose a solution which has the advantage to share the workload of harmonic filtering
between several active filter units. This solution was presented in [2] for the situation of
industrial power lines. In our situation we consider the case when the railway system is
equipped with active filter unit placed in the substation and locomotive has also an active
filter unit integrated. In this case using the existing control algorithms, the harmonic
filtering is not coordinated between the locomotive and substation because this would
required a real time communication between the AF unit of the locomotive and AF in the
substation. This impossible task can be performed using the distributed active filter solution
presented in [2] which has been proved to be valid in industrial power systems equipped
with distributed active filters (DAFS).
Using this method, several active filter units are installed along the electric power line. All
the active filter units operate as a harmonic conductance to reduce voltage harmonics. The
active filter unit acts as given:

, x x x h
i G U = (20)

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 31
where Ux,h represents the harmonic components of the line voltage Ux.
The line voltage Ux is measured and transformed into
e
xq
U and
e
xd
U using high pass filters
(HPF), the ripples of
e
xq
U and
e
xd
U are extracted. The voltage harmonics
e
xq
U and
e
xd
U are
then multiplied by the conductance command of the active filter. Based on the current
command and the measured current, the current regulator calculate the voltage command

( )
* * x
x x x x
L
v i i U
T
= +
A
(21)
In order to share the harmonic workload among the active filter units, it is necessary to use
the droop relationship between the conductance command and the volt-ampere of the active
filter unit.

0 0
( )
x x x x x
G G b S S = + (22)
where: Gx are the conductance command for the active filter units; G0 and S0 are the rated
operation point of each active filter unit.
The conductance command Gx of active filter unit (AFUx) is determinate by the volt-ampere
consumption of this active filter unit. To obtain the volt-ampere Sx of AFUx, the RMS values
of voltage and current associated with AFUx are calculated:

( ) ( )
( ) ( )
2 2
2 2
s s
xRMS xq xd
s s
xRMS xq xd
x xRMS xRMS
U U U dc
i i i dc
S U i

= +
`
)

= +
`
)
=
(23)
where the dc values are extracted by low-pass filters. Ex is the stationary frame value of the
current ix. We consider the case of coordination between the substation active filter AFUx
and railway locomotive filter AFUy. The volt-ampere associated with AFUx and AFUy can be
expressed as following:

, , ,
, , ,
x x x x h x f x x h
y y y y h y f y y h
S U G U U G U
S U G U U G U
= ~
= ~
(24)
The droop characteristics of both active filter units are given:

0 0
0 0
( )
( )
x x x x x
y y y y y
G G b S S
G G b S S
= +
= +
(25)
Based on (24) and (25), the relationship between Sx and Bx can be derived:

Fuzzy Controllers Recent Advances in Theory and Applications 32

0 0
0 0

y y y
x x x
x y
x y
b S G
b S G
S S
b b

= = (26)
If the droop characteristics of the active filter units are assigned as follows:

0 0 0 0 x x x y y y
G b S G b S = (27)

x x y y
b S b S ~ (28)
The slope of the droop should be set in inverse-proportion to the volt-ampere rating of each
AFUs to achieve the desired load distribution. This can be extended to the case of multiple
installations of active filter units.

0 0 x x y y
b S b S = (29)
The above droop settings allow harmonic filtering workload being shared in proportion to
the volt-ampere rating of the active filter units. The control structure of the proposed system
is presented in fig. 11. In our case since the system has a rather variable structure (due to the
traffic variability and locomotive distribution) we propose a fuzzy logic conductance
calculator.

Figure 11. Fuzzy logic control for the active filter distribution
This calculator identifies the appropriate value of the Gx according to the actual load and
load capability of the filter and the harmonic spectrum characteristic. Considering the case

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 33
when the active filter AF1 is located in substation and we have also an active filter operating
at locomotive level we can coordinate them using a load sharing algorithm which is
implemented by the fuzzy logic conductance command block.
The phase line voltage is measured and transformed into Uxq and Uxd in synchronous
reference frame. Using high pass filters (HPF), the voltage line harmonics are extracted and
multiplied by the conductance command Gx computed by the fuzzy logic conductance
controller. The current commands Ixs and Ixd of the active filter computed by the Clark
transformation block are transformed to the current commands Ix. Based on the current
command, the current regulator calculates the voltage command Vx as given:

* *
( )
x
x x x x
L
v i i U
T
= +
A
(30)
The fuzzy logic command can be tuned according to the capability of each unit in order to
have a different response for each situation.
The principle of operation of the fuzzy logic conductance command block is based on the
analysis of three parameters : the volt ampere Sx for the corresponding active filter, the total
THD of the current measured in active filter location and the averaged value of the sensitive
harmonics (in our case the 13
th
-17
th
harmonic).
1.4. Fuzzy logic conductance command block
The fuzzy logic conductance block is designed in order to reduce the harmonic distortion
components considering the distribution of the active filters. The distortion components
bare reduced using the fuzzy inference, the fuzzy logic conductance reacts differently in
function of filter volt-ampere, THD value and in order to avoid resonance, the average value
of 15
th
-17
th
-19
th
harmonics.
The variables of the inference system were represented as membership functions and the
normalization of the each input variable was made in order to match the value of the inputs
0 1 range.
In fig.s 12, 13 and 14 are presented the membership functions for the input magnitudes, and
in fig. 15 is presented the membership functions for the output magnitude.

Figure 12. Membership functions for Sx

Fuzzy Controllers Recent Advances in Theory and Applications 34

Figure 13. Membership functions for THD

Figure 14. Membership functions for Avg Harmonics

Figure 15. Membership functions for Gx Command
The most important element of the fuzzy systems is the base of rules which implements the
relationship between the inputs and the output of the system. The response surfaces of the
system are presented in fig.s 16 and 17.

Figure 16. Surface Gx vs. THD and Sx

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 35






Figure 17. Surface Gx vs. THD and log(Avg-Harmonic)
1.5. Simulation results
The control algorithm is implemented in PSIM and the control in Matlab Simulink and
fuzzy logic toolbox (fig. 18). The proposed solution is applied for the railway power
distribution system in order to demonstrate its capability to share the harmonic filter among
the various active filter units.
The current and voltage component on the filtered harmonic orders are reduced
dramatically and the typical effect of the 3
rd
and 5
th
harmonics have been neglected, but the
significant high frequency components (mainly 17
th
and 19
th
harmonics).
The circuit model of the railway is illustrated in fig. 19. Two active filter units are placed,
one in substation location and the 2
nd
at the locomotive level. The locomotive represents a
nonlinear load comprising several rectifier units. The parameters of the simulation are given
as follows: power system (27kV, 50Hz); line parameters usual values in railway
transportation 0,47/km.
The control structure is similar for each unit but the filters characteristics need to be tuned
according to their location.
Fig. 9 contains the wave forms when both active filter are working. The harmonic distortion
is improved as the THD are reduce by the filter in action of the AF1 and AF2.
The active filter units of the DAFS adjust the conductance command G1 and G2 based on
their own droop characteristics for the volt-ampere consumption for AF1 and AF2, which
indicates that the filtering workload is evenly shared between the active filter units of the
system (fig. 20 and 21).

Fuzzy Controllers Recent Advances in Theory and Applications 36

Figure 18. Fuzzy logic control for the distributed active filter

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 37

Figure 19. a) and b). Railway power system simulation

Figure 20. Variation of G1 and G2.

Figure 21. Variation of S1 and S2
(a) Substation model
(b) Locomotive model

Fuzzy Controllers Recent Advances in Theory and Applications 38
2. B. Advanced system for the control of work regime of railway electric
drive equipment
2.1. Introduction
Although the contact line in electric railway transport (and not only) can be regarded as a
line of electric power distribution, it has several (mechanical and electric) characteristics that
differ from the usual power systems. Thus, the most common failure regimes are the short
circuit in the contact line (L.C.), sub-sectioning (P.S.) or in traction sub-stations (ST),
produced by dielectric breakdown of insulators or mechanical failures caused by the
pantograph of the electric engine (fig. 22).
The sudden growth of currents in the case of short circuits has negative consequences both
upon the fixed installations of electric traction (thermal and dynamic effects) and upon other
nearby installations (dangerous inductions in the phone networks, in the low voltage grids
on pipelines, etc., which can endanger the life of the personnel in the vicinity of the railways.

Figure 22. Electric power feeding of railway traction
In the case of single-phase electric railways with the frequency of 50 Hz, the distance
between two ST is 50 60 km, depending on the profile of the line. For the simple lines, the
high value of characteristic impedance (0,47 O/km), leads to minimal short circuit currents
(at the end of the line), below the maximal load currents.
In the case of short circuits in the vicinity of the ST, the currents reach up to eight times the
charge current, and this is why they have to be cut off as quickly as possible. Neither the low
current short circuits must be maintained for too long, as they have negative effects upon
the installation.
Because of the elements mentioned above, protecting the contact line feeders by maximal
current protection is not efficient enough. At present, distance protection is being used,
either by di/dt relays or by impedance ones.
The di/dt relays function on the basis of the existing deforming regime existing in the
contact line, current regime which should be diminished as much as possible in the future.
The elimination of the deforming regime will make the di/dt relay inefficient.

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 39
Impedance relays, which are widely used in all power systems, can distinguish between an
overload and a short circuit [11]. Because of the configuration of the contact line, the shape
of the RX characteristic of the relay in the complex plane should be modified for each
traction substation separately, which is difficult to achieve in practice. In order that RX
characteristic meets the requirements imposed to the protection system, the complexity and
cost increase [12].
2.2. The structure of the protection device
In order to secure protection against the abnormal work regimes of the contact lines and
transformers, we designed a complex device meant to replace both the maximal current, low
voltage relays etc., and the impedance relay. The block diagram of the device is given in fig.
23 [10].

Figure 23. The block diagram of the protection device
The voltage of the power circuit is converted to a reduced value, accepted by the electronic
circuit, by means of a group of instrument transformers TR, thus achieving, at the same
time, galvanic separation of the power circuit using the insulation amplifier AI, which
works with unitary amplification factor. At the output of the insulation amplifier AI we
obtained an alternating voltage with a maximal amplitude of 10 V, fed both to a peak
detector DV1 and to a device meant to detect the passage through zero of the voltage wave,
DZU .
At the output of the peak detector DV1 we obtained signal V(U), whose value is
proportional to the value of the voltage in the power circuit. Signal V(U) is fed to one of the
analogue inputs of the fuzzy microcontroller. The voltage wave zero passage detector,

Fuzzy Controllers Recent Advances in Theory and Applications 40
marked DZU, is meant to generate an impulse at each such passage. The impulse obtained
at the output of this block is fed on the one hand to a monostable switch circuit CBM1, in
order to obtain a square impulse with a well determined shape and duration, and on the
other hand, to an initialization circuit, CINI1. The signal at the output of the monostable
switch circuit CBM1 is fed to the bistable switch circuit CBB1, triggering the switch of its
output at each impulse received. The bistable switch circuit CBB1 is meant to trigger the
determination the phase shift between the voltage and the current in the power circuit by
means of a validation/inhibition logic (block LVI) of the access of some square impulses
generated by the oscillator OSC 1 and fed to the asynchronous binary counter NB1.
The value of the current in the power circuit is converted by the ensemble current translator
amplifier TC+A into a voltage signal V(I) whose amplitude is 10V, which is fed to the peak
detector DV2. At the output of the peak detector we obtain signal V(I), which is fed to one of
the analogue inputs of the fuzzy controller. The current signal is also fed to a zero passage
detector of the signal corresponding to the current, DZI, which commands a monostable
switch circuit, CBM2, and a bistable switch circuit CBB2, both having similar functions with
those on the branch corresponding to the voltage wave. The bistable switch circuit CBB2 is
meant to stop the phase shift determination process, and it acts in this sense through the
validation/inhibition logic.
As shown before, we input to the fuzzy microcontroller three analogue signals V(U), V(I)
and V(), having values ranging between (05) V, whose values depend on the voltage,
current and phase shift in the power grid. According to the value of the three signals and
using the original processing program implemented on the fuzzy controller under the form
of processing rules, at its analogue output we obtained a continuous voltage ranging in the
domain (05)V. Its value is higher when the regime on the contact line approaches the
failure regime, respectively when it is under failure regime and it is directly proportional to
the seriousness of the failure.
What has to be done is to disconnect the contact line in a time interval that is inversely
proportional to the seriousness of the failure. At the same time, it is necessary to keep on
feeding the contact line in the situation of short term failures (caused for instance by
atmospheric overcharges).
Meeting these requirements can be achieved by a programmable time delay circuit. This is
made of a voltage frequency converter connected to the output of the fuzzy controller,
where a square signal is obtained, whose frequency is directly proportional to the voltage
given by the fuzzy controller. These square impulses are then fed to a binary counter NB3,
which shows during one counting period a value which is directly proportional to the
voltage at the output of the fuzzy microcontroller. Counter NB3 is set on 8 binary ranks. Its
outputs are fed to the input of a numeric comparator , CN. At the other inputs of the
numeric comparator we apply the outputs of another binary comparator NB2, that is meant
to memorize a value pre-established by the user. This value is input to the counter by means
of the adjusting block AJ, respectively of the tact oscillator OSC2. The impulses fed to binary
counter NB2 are simultaneously fed to a decimal code counter NBCD, whose outputs are

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 41
connected to a decoding system DEC and then to a 7-segment display, marked DISPLAY. In
this way, the user has a permanent control over the value given by binary counter NB2. This
value represents the very threshold that triggers the main contact line feeding interrupter.
Numeric comparator CN signals the situation when the numeric value in counter NB3
becomes equal to the threshold value in counter NB2. It is necessary to enable the
modification of the threshold value for the main interrupter, according to the various
normal functioning regimes established by the concrete practical situation and which have
to be dealt with accordingly. The threshold is established by the users, according to the
experience accumulated in time.
It is obvious that the triggering of the contact line main feeding interrupter is done in a time
interval that is directly proportional to the threshold value in binary counter NB2, and
inversely proportional to the voltage at the analogue output of the fuzzy microcontroller.
The main interrupter can be triggered almost instantaneously, if the fuzzy microcontroller
outputs a 5V voltage, which corresponds to highly dangerous failure. This value is sensed
by comparator C, which permanently compares the output value of the fuzzy controller to a
5V reference voltage, fed by reference source UR2.
In case of reaching a regime that triggers the main interrupter, at the output of port P we
have logic 1. This signal is power amplified by means of command signal amplifier ASC,
which outputs a high enough signal to command the main interrupter.
For a correct functioning, binary counter NB3 has to be periodically reset, so that its value
should not reach the threshold value memorized by binary counter NB2, even under a
normal functioning regime. This condition is met during each period of the voltage in the
grid, by initialization circuit CINI2. This circuit too, uses as time reference the zero passage
impulse of the grid voltage. It is preferable to initialize counter NB3 just once during one
period, as in this case, the value of counter NB3 depends on two operations of phase
determination corresponding to the two semiperiods which achieves in this way a high
immunity to very short perturbations or incorrect work regimes along the contact line.
All the circuits presented above have been designed built and tested and they worked correctly.
2.3. The Fuzzy Controller
As mentioned before, in order to offer a complex protection of the contact line in railway
electric transport, using mono-phase AC (27,5 [kV], 50 [Hz]) we designed a Fuzzy relay,
starting from the practical case of a railway electric traction sub-station [10]. This relay
allows a maximal current, overcharge, distance and minimal voltage protection. Although
initially we did not take into consideration directional protection (the inverse current
circulation between two traction substations), its introduction is not a problem and we
consider that the Fuzzy relay can achieve that too.
a. Information on the input magnitudes
We considered the following usual regimes:

Fuzzy Controllers Recent Advances in Theory and Applications 42
Variable
Linguistic
values
Variation
domain
Universe of
discourse [%]
I
n
p
u
t

Current
normal
50 600
[A]
030,8
overcharge
600800
[A]
25,641
short-
circuit
8002000
[A]
35,9100
Voltage
short-
circuit
16 20
[kV]
043,5
overcharge
20 25
[kV]
26,187
normal
25 27,5
[kV]
69,6100
Phase
shift
normal
030
[grad]
043,8
overcharge
3060
[grad]
31,381,3
short-
circuit
6080
[grad]
68,8100
O
u
t
p
u
t

Command
blocked 0 [V] 0
high delay 1,66 [V] 33,3
low delay 3,33 [V] 66,6
instant 5 [V] 100
Table 1. Input and Output Magnitudes
In figure 24 we present the recorded values for the current in one ST with high traffic
protected with usual distance relay. One may notice a rather high frequency of critical work
regimes transitory (three short-circuits at long distance, without disconnection due to a
mechanical failure: 1, 2, 4; and three overloads: 3, 5 and 6 all in a period of twenty minutes).
The waveforms for three cases: normal, overload and short-circuit are presented in figures 25,
26 and 27. The phase shift between current and voltage for three situations are clearly
different which justify the selection of phase as an important factor in identification of critical
regimes. Another interesting characteristic is the smoothing of the current waveform during
the short-circuit situation (even in the case of long distance short-circuit) and the phase shift
increase which is produced by the change of the complex impedance of the system.
As the scale representing the transfer functions that correspond to the processed magnitudes
is a percentage one (0 - 100%), we needed to norm the discussion universe, by means of a 1st
degree polynomial function.
In fig. 28, 29, 30 are presented the membership functions for the input magnitudes, and in
fig. 31 is present the output magnitude.

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 43

Figure 24. ST load current in different work conditions: 1,2,4 long distance short-circuit; 3,5,6
overload; 7 normal load

Figure 25. ST load current and voltage wave in normal regime

Figure 26. ST load current and voltage wave form in overload regime

Figure 27. ST load current and voltage in short-circuit regime

Figure 28. Representation of membership functions for the input magnitude current

Fuzzy Controllers Recent Advances in Theory and Applications 44


Figure 29. Representation of membership functions for the input magnitude voltage


Figure 30. Representation of membership functions for the input magnitude phase shift


Figure 31. Representation of membership functions for the output magnitude command
b. Command rules (inference)
The rules have been established for practical reasons after having checked the reference
literature and consulted experts in electric traction and protection installations for electrical
systems.
Fig. 32 shows the rules base which connecting the fuzzy input variables to the fuzzy
output variable, by means of the inference method (max/min). For deffuzification we
chose the Singleton weight centers method due to its major advantage, namely the short
processing time, a stringent condition for the real time functioning of the Fuzzy controller
with a function of relay. For this reason, for the practical application under analysis we
established Singleton-type membership functions corresponding to the linguistic term
command of the output magnitude. Using the max-min inference method alongside
with the defuzzification method is widely spread in practice and has lead to outstanding
performances of the regulation systems.

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 45
Fig. 33 shows the command surface for three constant values voltage corresponding to the
cases described above (normal, overload and short-circuit).

Figure 32. Rules base

Figure 33. Command surfaces for three voltage values

Fuzzy Controllers Recent Advances in Theory and Applications 46

Figure 34. Simulation diagram
Fuzzy controller simulation is realized in Matlab Simulink (fig. 34). We randomly
generated several combinations of input variables (fig. 35, 36 and 37), and the analysis of the
response (fig. 38) proved the correct functioning of the fuzzy system, according to the rules
base established.

Figure 35. Input magnitude CURRENT

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 47

Figure 36. Input magnitude VOLTAGE

Figure 37. Input magnitude PHASE SHIFT

Figure 38. Output magnitude COMMAND

Fuzzy Controllers Recent Advances in Theory and Applications 48
The fuzzy system has been implemented on an eZdsp TMS320F2812 development board
produced by Spectrum Digital.
2.4. Hardware implementation and testing for the Fuzzy Logic relay for railway
protection
The development cycle is based on three stages: first the fuzzy logic relay is designed and
tuned on Matlab Simulink using a Simpower System model of the railway ST because the
testing in the real system would have been extremely expensive and dangerous. This allows
the simulation, parameters tuning and optimization of rules data base.
In the second stage, the final structure of the fuzzy logic operations are prepared for
hardware implementation. In order to achieve a high speed operation some hardware
specific features of the DSP are employed to reduce the computation resources for the
application. One operation which requires intensive computation is the deffuzification
which need complex instruction. Each input (voltage, current and phase shift) is
sampled using an 8 bit ADC. The input variables are quantified in order to obtain the
degree of membership of each point of the input for each fuzzy set. For each point is
associated an index value corresponding to the membership function. The rule base in
coded using the same index system in order to have the correspondence between the
index of active membership functions with the corresponding part of the rule base.

Figure 39. Fault processing by the fuzzy logic relay
Finally, the structure of the system operation is coded using Code Composer Studio
development environment in C. The resulting assembly code is optimized for speed in order
a) long distance short circuit
b) relay response

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 49
to have the shortest cycle time. For the hardware implementation we have considered a
Spectrum Digital DSP development board with TMS320F2812 processor.
From the point of computation resources this can offer sufficient processing speed in order
to meet the application requirement at a reasonable development time (since no assembly
coding was necessary). This hardware solution used to implement the fuzzy logic relay
structure is simple and provide good performance/price output.
After manual verification of the correct functioning of the DSP fuzzy processor the entire
system was put under test in a real ST (in parallel with the existing protective system). In
figure 10 a one long distance short circuit it has been identified and the actual response of
the relay is presented in fig 10 b. One may notice the extremely fast response of the
proposed fuzzy relay system (the short circuit was not identified by the usual impedance
relay due to the misclassification of the fault).
The existing protection system obviously has considered the event as an overload and it
disconnected the main switch about one minute after.
3. C. Control system for catenary Pantograph dynamic interaction force
3.1. Introduction
An issue of great importance in railway electric transportation is the quality of the sliding
contact pantograph- catenary suspension.
An improper contact produces electric arcs with unfavorable consequences on energy loss,
reliability and wear of the subassemblies subjected to the arc, electromagnetic pollution, etc.
Of main interest is the behavior of the dynamic response of the two mechanical subsystems
coupled by the contact force (pantograph- catenary suspension), each of them having totally
different structures and dynamic properties.
The aims of the numerical simulation analysis were to obtain conclusive information for
kineto-static and dynamic characterization of the pantograph - catenary suspension
assemblys behavior.
Based on the simulations performed in this work, using an original program, was found that
the dynamic response of the studied assembly produces variation of the contact force in
very large ranges, depending on the trains travel speed and other parameters, being the one
which leads to detachments, thus to electric arcs.
Currently, there are several world-wide used models for pantographcatenary interaction,
which apply especially at high speeds (over 350km/h) and rely on different principles [19], [20],
[21]. These models have different degrees of complexity. Most of them consider simplifications,
as taking into account all the factors (i.e. aerodynamic forces, friction, proper oscillation of the
locomotive, etc.) leads to a very difficult problem. The traffic at speeds above 350km/h is made
on the specific lines, namely the railways, catenary suspensions and locomotives are especially
built and designed for these conditions. However, in Romania and the neighboring countries

Fuzzy Controllers Recent Advances in Theory and Applications 50
this traffic runs on normal lines, which through modernization can be improved to support
speeds from 140km/h up to 220km/h. In this case the railway, the catenary suspension and the
locomotives are just the upgraded classical ones. The current paper refers to this part of this
problem, which has not been extensively treated in the literature.
The conclusions presented in this work have a special importance, based on which
following to be achieved an intelligent management system of the pantographs pressure
force regime, taking into account the speed, its momentary position and the coupled
dynamic model of the two subassemblies: pantograph- catenary suspension.
3.2. Analysis of the Catenary-Pantograph assemblys dynamic behavior by
numerical simulation
The purpose of the numerical simulation, by results interpretation, was the possibility to test
the validity of the issued hypothesis upon the real systems behavior, being in operation, as
dynamically coupled subassemblies.
The analysis was made using the TENSI-CABLE calculation program, specially conceived by
one of this works authors for the dynamic behaviors study of some strength structures
composed by flexible elements.
The programs essence consists in the possibility to determine the structures response to a
dynamic stimulus, of excitation force or imposed motion type, into a section required
by the user. From space and paternity considerate of the calculation program, in this work
are presented and commented only the results of some of the performed simulations.
The study of the catenary suspensionpantograph assemblys dynamic behavior,
conducted by numerical simulation, was made considering a catenary suspension with
standard configuration in Romania, having the parameters written in table 2.

Material: Contact wire
Carrier cable
Articulated pendulum
- copper, section 100 mm
2
- galvanized steel, section 70mm
2

- galvanized wire, section 28 mm
2

Suspensions total length - 196 m
Distance between the supporting pillars - 3 spans of 60,0m each
Distance between the articulated pendulums - 6 m / 8m ( 9 pendulums per span)
Mechanical stress:
Upper wire
Lower wire
- 12.000N
- 12.000N
Specific mass:
Upper wire
Lower wire
- 0,89 kg/m
- 0,61 kg/m
Axial rigidity Upper wire
Lower wire
- 10500 kN
- 5000 kN
Table 2. Standards catenary parameters

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 51
The catenary simulated is a simple style catenary, which was chosen because it has all the
characteristic dynamic effects. A diagram of the model is given in fig. 40, where:
Tower stiffness: S; Dropper stiffness: K; Distance to the i-th tower: Wj; Distance to the i-th
dropper: Xj; Stiffness of the two wires: EIA; EIB; Density of the two wires: A; B; Tension in
the two wires: TA; TB.
The contact force, as excitation force considered in simulation, is the one afferent to a basic
pantograph of which designing parameters are typical to EP3 pantographs, being in current
operation at Romanian Railways, (fig. 40).
For the study made on the simulation model, it was built the input data block comprising
the determinant geometrical and mechanical parameters for the component elements of the
studied assembly.
In simulation were considered three spans of catenary (fig. 41), having a number of 70 nodes
and 103 elements.
The equations governing the response of the catenary are obtained through the
displacement of the contact wire and messenger wire, expressed as Fourier sine series
expansions. The displacement shapes of the contact wire and messenger wire from
equilibrium are each expressed as Fourier sine series expansions, and a good approximation
of the shape is possible if enough terms are used the series expansions are given below:

( )
, ( )sin
A m
m x
y x t A t
L
t | |
=
|
\ .

- messenger wire (31)

( )
, ( )sin
B m
m x
y x t B t
L
t | |
=
|
\ .

- contact wire (32)

where: y the wire displacement; Am the amplitude of the m-th sine term for the
messenger wire; Bm - the amplitude of the m-th sine term for the contact wire; x the
displacement along the catenary; L the total length of the catenary; m an integer,
designates the harmonic number.


Figure 40. Catenary model

Fuzzy Controllers Recent Advances in Theory and Applications 52

Figure 41. Pantograph model
The equations of motion contain displacement and acceleration terms ( , , , A A B B
-- --
). When the
system is excited, the catenarys response is harmonic so the acceleration terms are:

2
m
m
A A e
--
= (33)

2
m
m
B B e
--
= (34)
where: e - the natural frequency of vibration
The orthogonal modes of the catenary can be considered separately and the result from
modal analysis gives the equation for each mode:

2
2
i i
i i i i i i i i
M z M z MZ Q e e
-- -
+ + = (35)
where: zi(t) the i-th model response; mi the i-th modal mass;
i
- the damping ratio; ei
the i-th natural frequency; Qi the i-th modal forcing function.
Simulation started at time t=0 seconds with the pantograph at pillar zero, on the contact wire
acting the pantographs lifting force, becoming contact force starting with this moment.
By pantographs forwarding along the running track, it moves successively the catenarys
sections upwards and, thus, induces a vibrating motion of the wire of which oscillations,
after pantographs passing, propagate along the contact wire.

Figure 42. The three spans of the catenary

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 53

Figure 43. Propagation of the oscillating motion induced by the pantographs excitation force and
catenarys deformation, after t=8,98s

Figure 44. Contact force at different speed
Exemplifying of this phenomenon is shown in fig. 43 that includes also the deformed shape
of the catenary suspension, afferent to the pantographs position right at a node located in
the last third of the first span.
The shape shifting of the catenarys distortion right at the pillars (from a steep descent to a
slope less pronounced) has the most pronounced effect on the pantographs performances,
the operation data confirming also that in these areas are noticed also the most frequent

Fuzzy Controllers Recent Advances in Theory and Applications 54
contact losses. This conclusion is confirmed also in the specialty literature, the authors
finding that at speeds over 150km/h, even the pantographs crosshead shoe could keep the
contact (because the upper suspension, being lighter, could accommodate to a greater
movement between the crosshead shoe and the frame), the pantographs frame responds
slower (because, having a greater mass and inertia forces are higher) and, while the
pantograph passes the area in the right of a supporting pillar, it sub-vibrates, the result
being a smaller contact force.
Table 3 shows the characteristics of contact force at a different speed. It can be seen that with
the increasing of speed the characteristics of contact force are getting worse. The speed of
160 km/h is a critical point. At this speed, the contact force approaches to zero. It means that
the slipper-shoe of the pantograph lost contact with the overhead line.

Speed (km/h) 100 120 140 160 170 180
Max. contact force (N)
Max. contact force (N)
99
42
103
40
105
24
117
2
119
0
367
0
Table 3. The contact force at different speed
When the running speed is less than this speed, the contact force is larger than zero and the
pantograph catenary system works well. But when the running speed is larger than 160
km/h, loss of contact occurs and becomes more frequent with the increasing of the speed.
Considering the minimum contact force should be larger than 20 ~ 30 N in order to keep
good contact, we can see from the result that the EP3 pantograph together with the tested
catenary, is only suitable for running speed less than 140 km/h. Figure 44 shows the contact
force under a speed of 140, 160 and 180 km/h. At a speed of 180 km/h, the contact is lost, and
the impact causes large maximum contact force.
3.3. Active control strategies
Studying the previously presented problem, one can observe that the contact force between
the pantograph and the catenary is variable, due to a different dynamical behavior of the
two parts, with maxima and minima which occur at a frequency of 0.52 [Hz].
For flow speeds over 350km/h, on special lines built and equipped for this purpose, one can
find in the literature active control solutions of contact force through "classical" methods [22]
[23] [24] [25] [26] or advanced methods [27], [28], [29].
These are quite expensive and not profitable when used at lower speeds (160km/h
220km/h). For this case, we propose a contact force control system of the pantograph-
catenary based on the fuzzy logic.
Our aim is to reduce variations in the contact force, by eliminating the pantograph
detachment and the appearance of excessive peaks (Table 3), which lead to a premature
wear of the contact wire. The structure of the contact force control system is shown in
Figure 45.

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 55

Figure 45. System structure

Figure 46. Fuzzy controller
The proposed fuzzy controller presented in Figure 46, will determine the correction through
which the contact force will be closer to the desired value, using the information given by
the error (c) and rate of change of the error (Ac).
The variables (c) and (Ac) are converted into fuzzy variables, using membership functions
and the discussion universes in Figures 47, 48 and 49.
The variables (c) and (Ac) are processed using the inference of the 49 rules (Table 4) through
the method of min/max. Output defuzification is made by the centroid method.

Figure 47. Membership functions for error

Fuzzy Controllers Recent Advances in Theory and Applications 56

Figure 48. Membership functions for speed error

Figure 49. Membership functions for command
Table 4. Rule base
The control area of the proposed fuzzy controller is shown in Figure 50.
Using a Matlab-Simulink simulation, we performed a fuzzy controller functionality
simulation. The simulation results are presented in Figures 51, 52 and 53.
For the signal error, a triangular variation was chosen, which went through all of the
conversation universe (with a linear increase error from -0.1 to +0.1, followed by a linear
decrease to the initial value). The derivative of the error has constant maximum positive
values within the first ten seconds and minimum negative values in the next ten seconds.

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 57
The command variable follows the base of rules, namely while the error grows, the
correction of contact force gets stronger, as there is a non-linear relationship between the
two variables, which is in agreement with the experiment.

Figure 50. Control surface

Figure 51. Input signal for error

Figure 52. Input signal for speed error

Fuzzy Controllers Recent Advances in Theory and Applications 58

Figure 53. Output signal
When the error changes sign, the strength correction follows the same principle, but in the
opposite direction. One concludes that the established rules lead to contact force
stabilization. Using the original TENSI-CABLE model and applying the corrections
established by the fuzzy system over the dynamic contact force, we were able to estimate the
new variation form of the force between the pantograph and the catenary.
Note the small variations 10N around the value of 75N, which was considered to be the
optimal value of the contact force in order to achieve a correct transfer of energy.
The above mentioned variations may be reduced even more in the real implementation
phase of the controller, by using a correction of the base of rules.
The speed of response of the control system is very high: only the upper arms have to be
moved, reducing the mass to be moved with respect to the case of the force acting of the
linkage frame and moving the mass of the whole pantograph. A brushless DC motor with
low inertia and small electro-mechanical time constant could be the simplest choice for the
electrical actuator.
An important observation must be considered, independently upon the chosen control
strategy. In order to overcome the first drawback, the uplift contact force can be subdivised
into two terms: a mean term given by a constant force, provided by the traditional passive
mechanism (usually a spring or pneumatic cylinder) as in standard pantographs and an
additional oscillating term provided by a closed-loop active actuator. In such a way the
current collection capability can be maintained even in a case of a failure in the active
control apparatus.
A crucial problem to be overcome for a practical implementation of closed-loops active
pantograph is the measurement of the contact force. The force sensor must be reliable and
lasting in a harsh environment. The traditional solution is to measure the contact force at the
support points of the contact strips and to compensate for the inertia of the contact strip by
an accelerometer. The required instrumentation is expensive and its life-time is short due to
the high temperature and to the vibrational phenomene. Ref [23] presents other methods of
simulation for the contact force, which eliminate the previously mentioned disadvantages.

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 59

Figure 54. Contact force corrected
4. Conclusion
It can be concluded that usual active harmonic filtering the network side harmonic
distortion is reduced very effectively, but the psophometric effect caused by the current
along the traction system basically did not change [9]. This situation cannot be tolerated
since the effects of induced currents due to the harmonic components and resonance can
produce catastrophic effects upon the various equipments placed in the vicinity of the
railway power lines.
The proposed Fuzzy DAFS can deploy several small-rated active filter units at various
locations within the facility to damp the harmonic resonance. Compared to a centralized,
large rated active filter, the distributed, small-rated active filters with fuzzy logic
controlled DAFS can reduce harmonics in the railway power system with improved
effectiveness.
The advantage of the presented system based on fuzzy logic is the fact the use of a rather
distributed system which is capable to adapt to the railway configuration. Using several
filter units, the system has the capability to eliminate harmonic pollution and to avoid
harmonic resonance in railway systems. By means using simulation results, we have
validating the system considering the coordination of two active filters: one placed in
substation area, and the 2
nd
placed at the end of the line (locomotive level).
Part (B) introduces a complex for the protection of the contact line and of the railway electric
transport installations, based on the fuzzy logic. The system we presented has been built in
practice and its experimental implementation has shown a series of advantages with respect
to the actual systems, such as: low cost; adaptability to any real situation by the appropriate
modification of the rules base; very short response time (5s) of the fuzzy controller; it
functionally replaces several protection systems, taking over all their functions. The system
has the great advantage over other fuzzy logic relay systems that is designed considering
the specific railway power system behavior.
Based on an original numerical model for the dynamical interaction pantograph-catenary,
part (C) analyzes the evolution of the contact force between them, concluding that control

Fuzzy Controllers Recent Advances in Theory and Applications 60
could be lost at speeds above 170km/h, with multiple disadvantages. The mathematical
model is complex, nonlinear and can not be used directly in managing online the contact
force regime. This paper proposes a simple solution based on the fuzzy logic, which
eliminates excessive variations, therefore enabling a good electrical contact between the
pantograph and the catenary. The system was designed for the traffic of electric locomotives
on classical lines, with speeds up to 220 km/h.
Author details
Stela Rusu-Anghel and Lucian Gherman
Politehnica University of Timisoara, Romania
5. References
[1] P. Kiss
1
, A. Balogh
2
, A. Dan
1
, I. Varjasi
2
, The Application of Active Filter Supported by
Pulse Width Modulated Inverters in the Harmonic Simulation of the High Power
Electric Traction , Proceedings of ICREPQ 2008 Canary Islands, pp. 231-236.
[2] Po-Tai Cheng, Zhung-Lin Lee, Distributed Active Filter Systems (DAFS): A new
approach to power system harmonics, The 39
th
IEEE Conference IAS 2004, Seattle, USA,
pp. 94 101.
[3] H. Akagi, Modern active filters and traditional passive filters, Bulletin of the Polish
Academy of Sciences Technical Sciences, Vol. 54, No. 3, 2006, Warsaw , pp. 255 269.
[4] B. Dobrucky, M. Pokorny, V. Racek, R. Havrila, A New Method of the Instantaneous
Reactive Power Determination for Single-Phase Power Electronics Systems, Proceedings
EPE 2003 Lausanne , pp. 1 9.
[5] Juraj Altus
1)
, Jan Michalik
1)
, Branislav Dobrucky
1)
, L.H. Viet
2)
, Single-Phase Power Active
Filter Using Instantaneous Reactive Power Theory Theoretical and Practical
Approach, Journal of Electrical Power Quality and Utilization, Vol. 11, No. 1, 2005, pp. 33
38.
[6] Ahmed A. Helal, Nahla E. Zakzouk, and Yasser G. Desouky, Fuzzy Logic Controlled
Shunt Active Power Filter for Three-phase Four-wire Systems with Balanced and
Unbalanced Loads, Proceedings World Academy of science, Engineering and Technology 58
2009, ISSN 2070 - 3740, pp. 621 627.
[7] Soumia Kerrouche, Fateh Krim, Three-phase Active Filter Based on Fuzzy Logic
Controller, International Journal of Sciences and Techniques of Automatic Control & computer
engineering IJ-STA, Vol. 3, No. 1, 2009, pp. 942 955.
[8] Victor M. Moreno
1
, Alberto Pigazo
1
, Improved Injection Current Controller in Single-
Phase Shunt Active Power Filters, International Conference on Renewable Energy and Power
Quality ICREPQ 2005, Saragossa, Spain
[9] Hideaki Fujita, Takahiro Yamasaki, Hirofumi Akagi, A Hybrid Active Filter for Damping
of Harmonic resonance in Industrial Power Systems, Proceedings IEEE Transactionson
Power Electronics, Vol.15, No.2, March 2000, Japan, pp. 215 222.

Embedded Fuzzy Logic Controllers in Electric Railway Transportation Systems 61
[10] S. Rusu Anghel and others, Advanced System for the Control of Work Regime of
Railway Electric Drive Equipment, Contract research no. 53/2009, Politehnica
University of Timisoara SC. SOFTRONIC S.A. Craiova, Romania, 2009.
[11] Popovic, T. Kuhn, M. Automated fault analysis: From requirements to
implementation, Power & energy Society General Meeting, 2009. PES 09. IEEE, pp 1
6, 2009, Calgary, AB.
[12] Saleem, A.Z. ; Khan, Z.A. ; Imran, A. Algorithms and hardware design of modern
numeric overcurrent and distance relays, Electrical Engineering, 2008. ICEE 2008.
Second International Conference on IEEE, pp 1 5, 2008, Lahore.
[13] Soliman, W.M. ; Soudy, B.E.-D.H. ; Wahab, M.A.A. ; Mansour, M.M. Power generation
station faults diagnosis based on fuzzy relations using information of protective relays
and circuit breakers, Electric Power and Energy Conversion Systems, 2009. EPECS 09.
International Conference on IEEE, pp 1 6, 2010, Sharjah.
[14] Farzad Zhalefar, Majid Sanaye-Pasand, A new Fuzzy-logic-based Extendet Blocking
Scheme for Differential Protection of power Transformers, Control and Intelligent
Processing Center of Excellence, School of Electrical and Computer Engineering,
University of Tehran, Tehran, Iran, Eds. Taylor & Francis, vol 38, pp 675-694, 2010.
[15] Hubertus, J ; Mooney, J ; Alexander, G. Application Considerations for Distance Relays
on Impedance-Grounded Systems, Protective Relay Engineers, 2008 61st Annual
Conference for IEEE, pp 196-203, 2008, College Station, TX.
[16] Youssef, O.A.S. Applications of fuzzy inference mechanisms to power system
relaying Power Systems Conference and Exposition, 2004. IEEE PES, vol 1, pp 560
567, 2005.
[17] Spectrum Digital Inc, Application Note The Design and Development of Fuzzy Logic
Controllers, 2009.
[18] N. Rusu, J. Averseng, C. Miklos, C. Alic, S. Anghel Dynamic Modeling of Pantograph
Catenary System for Energy Loss Control Conference on Automation, Quality and
Testing, Proceedings of AQTR-IEEE, vol III pp 11 14, Cluj-Napoca, Romania, 2006.
[19] A. Carnicero, J. Simenez, A. Ramos, C. Sanchez Simulation Models of the Catenary
Pantograph Dynamic Interaction: Validations and Applications , Proceedings of
Power Transmissions in High Speed Railways Systems Seminar, pp 13 26, ESIEE
Awiens, France, 2009.
[20] J. Bennet et al, Advanced Algorithm to Calculate Mechanical Forces on a Catenary,
Computers in Railways IX, pp 857 868, Germany, 05.2009.
[21] Weihua Zhang et al, Hybrid Simulation of Dynamics for the Pantograph Catenary
System, Vehicle System Dynamics, Volume 38, Issue 6, Pp 393 414, 2002.
[22] M. A. Abdullah, Y. Michitsuji, M. Nagai, N. Miyajima, Integrated Simulation between
Flexible Body of Catenary and Active Control Pantograph for Contact force Variation
Control, Journal of Mechanical Systems for Transportation and Logistics, Vol 3, No. 1,
2010 pp 166 177 .
[23] G. Diana, F. Fossati, F. Resta, High Speed Railway: Collecting Pantographs Active
control and Overhead Lines Diagnostic Solutions, Vehicle System Dynamics, Volume
30, Issue 1, pp 69 84, 1998.

Fuzzy Controllers Recent Advances in Theory and Applications 62
[24] D.N. OConnor, S.D. Eppinger, W.P. Seering, D.N. Wormley, Active Control of a High-
Speed Pantograph, Journal of Dynamic Systems, Measurement, and Control, Vol 119,
pp 1 - 4, March 1997
[25] A. Balestrino, O. Bruno, A. Landi, L. Sani, Innovative Solutions for Overhead
Catenary-Pantograph System : Wire Actuated control and Observed Contact Force,
Vehicle System Dynamics, Volume 33, Issue 2, pp 69 89, 2000.
[26] A. Levant, A. Pisano, E. Usai, Output-Feedback Control of the Contact Force in High-
Speed-Train Pantographs, Proceedings of the 40
th
IEEE Conference on Decision and
Control, pp 1831 1836, Orlando, Florida USA, December 2001.
[27] B. Allotta, A. Pisano, L. Pugi, E. Usai, VSC of servo-actuated ATR90-type pantograph,
Proceedings of the 44
th
IEEE Conference on Decision and the european Control
Conference 2005, pp 590 - 595, Seville, Spain, December 2005.
[28] Y. J. Huang, Discrete fuzzy variable structure control for pantograph position control,
Electrical Engineering 2004, No 86, pp 171 177.
[29] Y. J. Huang, T. C. Kuo, Discrete Pantograph Position Control for the High Speed
Transportation Systems, Proceedings of the 2004 IEEE International Conference on
Networking, Sensing & Control Taipei, Taiwan, pp 21 23, March, 2004.
Chapter 3
Design of a Real Coded GA
Based Fuzzy Controller for Speed
Control of a Brushless DC Motor
Omer Aydogdu and Ramazan Akkaya
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/49972
1. Introduction
Fuzzy controllers are nonlinear elements used in the control of linguistically defined
systems, which can not be modeled accurately. Besides, they are an effective approach for
various complex and ill-defined systems [1,2]. In design of fuzzy controllers, there is not
well defined approach. The sophisticated and tedious design process is usually
implemented by an expert. The success of the controller depends on the knowledge and skill
of the expert. In some cases, even a very experienced and skillful experts extensive efforts
may not yield optimal solution for fuzzy controller design. The optimal design of fuzzy
controller has a critical role for their more widespread and effective use. The design
inherently requires the determination of a great deal of features and parameters. For this
reason, fuzzy controller design problem has a number of local values in a large solution
space in the direction of a number of objectives. The conventional trial-and-error based
method makes the solution of the problem very difficult [3].
Genetic algorithms (GAs) are optimization strategies performing a stochastic search
by iteratively processing populations of solutions according to their fitness, i.e. a
predefined scalar index of satisfaction of the design objectives. In control applications, the
fitness is usually related to performance measures such as integral error, settling time, and
so on. Fitness function may contain more than one objective like the minimization of settling
time, steady state error and maximum overshoot. Thus, fitness function can be addressed as
a multi objective function. GAs are effective in solving multi-objective optimization
problems [4,5].
Most of the GA approaches in use represent the constraint variable using binary form of
coding. One of the major disadvantages of using binary coding is the slow convergence

Fuzzy Controllers Recent Advances in Theory and Applications 64
speed of the fitness function. Binary coding is also not at all efficient to be used in
computer memory. Therefore, the use of real coded genetic algorithm can overcome the
inefficient use of computer memory and can contribute the performance. This
contribution becomes clear when a lot of parameters are needed to be adjusted in the
same problem and higher precision is required for the final result. In literature, for real
valued numerical optimization problem, floating point representations were proven to
outperform binary representations because they are more consistent, more precise and
they lead to faster execution. Also, real coded GA is inherently faster than the binary code
GA, because the chromosomes do not have to be decoded prior to the evaluation of the
objective function [6].
Brushless DC (BLDC) motors are one of the motor types rapidly gaining popularity. Due to
their favorable electrical and mechanical properties like high starting torque, high efficiency
and noiseless operation, the BLDC motors are widely used in various consumer and
industrial systems such as actuation, robotics, machine tools, servo motor drives, home
appliances, computer peripherals, and automotive applications [5,7]. Operation of the BLDC
motors requires non-linear control due to their non-linear characteristics and presence of
sensors to estimate rotor position. Moreover, owing to some drawbacks of position sensors
such as cost, space requirement and instability, sensorless speed control has recently gained
importance [8]. It is possible to determine when to commutate the motor drive voltages by
sensing the back emf voltage on an undriven motor terminal during one of the drive phases.
Use of fuzzy controllers that have non-linear processes such as fuzzification, defuzzification
and fuzzy inference is suitable for BLDC motor control.
In this study, sensorless speed control of the BLDC motor with real coded GA based fuzzy
controller has been designed, simulated and practically implemented. ADSP-21992 digital
signal processor (DSP) is used to realize the conventional and optimal fuzzy controller
algorithms and sensorless speed control of the BLDC motor has been experimentally
implemented successfully.
2. Real coded GA based fuzzy controller
The realized control system general diagram is shown in Figure 1. As shown in the figure,
the control system includes two closed loops. The inner loop is the fuzzy controller loop that
accomplishes speed control of BLDC motor. The outer loop, which processes the fuzzy
controller and system operations in background, is the genetic algorithm loop that tunes the
controller parameters in regard to performance index of the control system.
2.1. Fuzzy controller
In this study, Mamdani type fuzzy controller, which has five blocks namely normalization,
fuzzifier, inference mechanism, defuzzifier and denormalization, has been used [8]. Block
diagram of the real coded GA based fuzzy controller consisting two inputs (e1, e2) and one
output (u) is shown in Figure 2.

Design of a Real Coded GA Based Fuzzy Controller for Speed Control of a Brushless DC Motor 65

Figure 1. Real Coded GA based fuzzy control of the BLDC motor

Figure 2. Block diagram of the Real Coded GA based fuzzy controller
In closed-loop control systems, the use of error (e1) and the change in error (e2) as controller
input is a universal approach. In the implemented fuzzy controller, error and change in
error have been used as inputs. As given in Eq. (1), the error is the difference between the
reference speed and the actual rotor speed.

1

*
e (t) (t) - (t) =
(1)
Here, e1(t) is the speed error,
*
(t) is the reference speed and (t) is the actual motor speed.
The change in error e2 (t) is determined by Eq. (2) [9,10].

2 1
=
d
e (t) e (t)
dt
(2)
In a fuzzy control system as shown in Figure 2, two normalization parameters (n1, n2) for
inputs (e1, e2) and one denormalization parameter (n3) for output (u) is defined. In

Fuzzy Controllers Recent Advances in Theory and Applications 66
normalization process, the input values are scaled in the range [-1,+1] and in
denormalization process the output values of fuzzy controller are converted to a value
depending on the terminal control element. Obtaining the normalization and
denormalization parameters of fuzzy controller is important for system stability.
In the fuzzifier process, the crisp input values (e1, e2) are converted into fuzzy values. Also,
the fuzzy values obtained in fuzzy inference mechanism must be converted to crisp output
value (u) by defuzzifier process. For this purpose, triangular fuzzy membership function is
defined for each input and output values by seven clusters. Figure 3, illustrates the
membership function used to fuzzify two input values (e1, e2) and defuziffy output (u) of the
realized fuzzy controller. For seven clusters in the membership functions as shown in Figure
3, seven linguistic variables are defined as; Negative Big (NB), Negative Medium (NM),
Negative Small (NS), Zero (Z), Positive Small (PS), Positive Medium (PM) and Positive Big
(PB). Initially, the overlap rates of membership functions are 50%. As illustrated in Figure 3,
peak or bottom points of membership functions to be tuned are defined as a1 and a2 for e1, b1
and b2 for e2, c1 and c2 for u. Therefore, the design of optimal fuzzy controller requires
optimization of at least six parameters (a1, a2, b1, b2, c1, c2) by using real coded GA for
fuzzification and defuzzification processes that are both nonlinear.

Figure 3. Membership functions of (a) input (e1,e2), (b) output (u)
In this study, the center-of-gravity method given by Eq. (3) has been used in defuzzifier
process for simulation and real time DSP application.

1
1
m
d A( )
i i
i
u
m
A( )
i
i

=
=

=
(3)
Here, u is the fuzzy controller output, di is the distance between i
th
fuzzy set and the center,
A(i) is the area value of i
th
fuzzy set.
The rule definition is subjective and based on the experts knowledge and experience. For
the system with two inputs and seven membership functions in each range, it leads to a 7x7
decision table and 49 fuzzy rules. For example, two fuzzy rules are decribed as;

Design of a Real Coded GA Based Fuzzy Controller for Speed Control of a Brushless DC Motor 67
if e1 is NB and e2 is NB then u is R1,
if e1 is NB and e2 is NM then u is R2,
Sliding mode rule base table used by fuzzy controller is given in Table 1. In the sliding mode
rule base, when an assumption is made such that R1= -R13, R2= -R12, R3= -R11, R4= -R10, R5= -R9,
R6= -R8, it is required to determine a minimum of seven parameters (R1-R7) by using real
coded GA.
Input-e1
Input-e2
NB NM NS Z PS PM PB
NB R1 R2 R3 R4 R5 R6 R7
NM R2 R3 R4 R5 R6 R7 R8
NS R3 R4 R5 R6 R7 R8 R9
Z R4 R5 R6 R7 R8 R9 R10
PS R5 R6 R7 R8 R9 R10 R11
PM R6 R7 R8 R9 R10 R11 R12
PB R7 R8 R9 R10 R11 R12 R13
Table 1. Sliding mode rule base
The developed fuzzy logic uses the min-max compositional rule of inference. The inference
mechanism of fuzzy controller is implemented in regard to the rule base given by Eq. (4).

1 2 i i i
u min e , e ( ) = ( ( ) ( ) )
(4)
Fuzzy controllers are nonlinear tools because of the nonlinearity of logical inference,
fuzzifier and defuzzifier processes. As mentioned earlier, design of fuzzy controller requires
determination of a minimum of 16 parameters comprised of three normalization
parameters, six membership function parameters and seven sliding mode rule base
parameters. The parameters are decided in regard to the symetrical properties of the
controller and that imposes limitations for the controller. The reduction of the number of
parameters gives acceptable results and it can be preferred in design process to simplify the
sophisticated optimization process.
2.2. Optimization of fuzzy controller by means of real coded GA
In problem solving, the precision of the numbers is crucial. In binary coded GA, the accuracy is
limited by the size of the chromosomes. However, the use of real coded GA, which can be
coded by real numbers, is advantageous. Real coded GA is more accurate and occupies less
space in memory. In literature, it has been reported that, real coded GAs operate faster than
binary coded GAs and they can converge to global optimum faster [11]. Also, in the
optimization of systems consisting a great deal of parameters to be optimized as in fuzzy
controllers, the chromosomes in the binary coded GA becomes too long and the parameter
accuracy can not be handled. However, in the developed algorithm, the parameters are coded
by integer number set and the parameter accuracy can be determined arbitrarily. Also, in the

Fuzzy Controllers Recent Advances in Theory and Applications 68
developed algorithm, binary coding can be accomplished in the limited valued parameters for
the optimization of sliding mode rule base table (R1, R2, R3, R4, R5, R6, R7).
In Figure 4, the flowchart of real coded GA used in the study is shown. Some of the GA
parameters such as the possible lower and upper limit values of parameters, the precision of
parameters, the termination criterion or loop number, the mutation probability, the crossover
probability, the population number and elitism property is required to be initialized.

Figure 4. Flowchart of the Real Coded GA based fuzzy controller
In the proposed algorithm, the normalization parameters (n1, n2) and de-normalization
parameter (n3), the membership function parameters (a1, a2, b1, b2, c1, c2) and rule base (R1, R2, R3,
R4, R5, R6, R7) of the fuzzy controller have been optimized. Two approaches have been used in
determination of parameters in optimal design of fuzzy controller by using real coded GA. In
the first approach, the parameters for optimal fuzzy controller have been determined

Design of a Real Coded GA Based Fuzzy Controller for Speed Control of a Brushless DC Motor 69
sequentially. This approach is similar to the self-tuning adaptive system employed in adaptive
control systems. The second approach is a method based on the simultaneous optimization of
normalization factors, membership functions and rules. Fuzzy controller parameters are not
fully independent of each other [12]. For this reason, simultaneously tuning of all the
parameters is the optimum solution. However, the large number of fuzzy control parameters
makes the second approach more difficult to implement.
The performance index defined for the fuzzy controller is given in Eq.(5) [3]. This equation
has also been used as multiple objective function employed in optimization process for real
coded GA.

1
1
1 1 1
0 0
6
t
t t
in
t
d
J e dt e t dt e dt
dt
| |
= + +
|
\ .
} } }
(5)
In Eq.(5), Jin is the performance index used as fitness value in real coded GA, e1 is the error
value and t1 is the settling time of the reference speed in BLDC motor fuzzy control system.
In Figure 5 (a), the coding of three normalization parameters in six digit number string in
real coded GA is depicted. Here, in six digit number string, two of it is assigned for integer
part and four of it is for the fractional part. As shown in Figure 5 (b), a total of six coding has
been implemented for three membership functions. Here, a1 and a2 are the membership
function parameters defined for input (e1); b1 and b2 are the membership function parameters
defined for input (e2); and c1 and c2 are the membership function parameters defined for
output (u). All obtained membership function parameters are less than one. In Figure 5 (c),
the sliding mode rule base parameters presented in Table 1 are shown. These parameters
take integer values in the range [1,2,3]. Therefore, each parameter is denoted by one digit.
As shown in Figure 5, a total of 63 digits have been used for all of the codes. Additionally,
the precision of parameters can be increased by increasing the number of digits. In coding
structure shown in Figure 5, the precision of membership functions is 0.0001.

Figure 5. Coding of the fuzzy parameters in real coded GA, (a) Normalization, (b) Membership
functions, and (c) Sliding mode rule table

Fuzzy Controllers Recent Advances in Theory and Applications 70
In the implemented algorithm, the initial population can be preferably set by the user or can
be randomly appointed a value in regard to Eq.(6). An important feature of random
generator is that in each execution of the algorithm different values are initialized.

ipop h pop
N (P - P ) Random (P ) P = +

(6)
Here, Nipop is the initial population, Ph and P are the minimum and maximum values the
parameter can have and Ppop is the random number generated between zero and one. Fitness
value of each chromosome in initial population is determined by using performance index
which is given by Eq.(5). As shown in the Figure 4, if the termination criterion is met, real
coded GA loop ends but optimal fuzzy control system keeps operating. Else, next step in
real coded GA is executed. In this study, Roulette wheel method is employed in natural
selection process. In binary coded GA, the crossover operation can be implemented by using
methods like single-point or two-point crossover. However, in real coded GA, the mixing
methods give better results in crossover operation. In mixing method, the values of two
parameter are compared and new generations are produced using Eq. (7).

1
new an bn
P P ( )P = +
(7)
where | is a number randomly generated between 0 and 1, Pan is the n
th
parameter of mother
chromosome, Pbn is the n
th
parameter of father chromosome [13,14].
Genetic algorithms sometimes converge prematurely. In order to prevent convergence to a
local point, new solutions are obtained by means of mutation. In this study, the number of
parameters to be applied to the mutation operation has been determined by multiplying the
total number of parameters with mutation rate. The elements to apply mutation operation
are determined and they are replaced by randomly generated numbers.
In the developed algorithm, the elitism, a mechanism in which the individuals with the best
fitness values in previous population are guaranteed their place in the next population, is
implemented by selecting. Later, new fitness value is determined again for new population.
Optimization process goes on until the termination criterion is met.
3. Sensorless control of BLDC motor
The BLDC control drive system is based on the feedback of rotor position, which is obtained
at fixed points typically every 60 electrical degrees for six-step commutations of the phase
currents. So, the operation of BLDC motor requires a control system and position sensors to
estimate rotor position. Rotor position is sensed using Hall effect sensors embedded into the
stator frequently. However, sensorless speed control has recently gained importance owing
to the elimination of some drawbacks of sensors such as cost, space requirement and
instability. Figure 6 shows basic block diagram of sensorless control of BLDC motor with
real coded GA based fuzzy controller.
In Figure 6,
*
is the reference speed (rad/sec), is the actual rotor speed (rad/sec), is the
rotor position (degree), u is the control signal used to reference moment (N-m), ia, ib, ic are

Design of a Real Coded GA Based Fuzzy Controller for Speed Control of a Brushless DC Motor 71
the actual phase currents (Amper),
*
a
i ,
*
b
i ,
*
c
i are the reference phase currents (Amper), S1-
S6 are switches of the inverter and Vdc is the supply voltage of the inverter (Volt).


Figure 6. Block diagram of sensorless control of the BLDC motor drive system
In speed control loop as shown in the block digram, the reference speed and the actual
motor speed is compared and the error signal is obtained. These signals (e1, e2) are employed
in fuzzy controller and reference current (I
*
) is produced for control system. The current
control loop regulates the BLDC motor current to the reference current value generated by
the speed controller. The current control loop consists of reference current generator, PWM
current control unit and a three phase voltage source inverter (VSI). Position of the BLDC
motor is obtained by employing zero crossing back emf detection method eliminating
position sensor requirement.
3.1. Modeling of the BLDC motor
Figure 7 describes the basic building blocks of the BLDC motor and inverter that results in a
system producing a linear speed-torque characteristic similar to the conventional DC motor.
BLDC motor has three phase windings on the stator similar to three phase squirrel cage
induction motor and magnets are placed on the rotor to provide air gap flux resulting in
brushless rotor construction. When the motor is operated at a certain speed, trapezoidal
emfs are induced in stator phase windings. The quasi-square wave AC current is fed to
stator phase windings through electronic commutator using current controlled voltage
source inverter and rotor position sensor resulting in constant torque development by the
motor.
At any instant, two out of three phase stator windings of the motor carry currents
synchronized with developed electromagnetic torque as shown in Figure 8. Active
switching states for three phase inverter operation, three phase back-emf waveforms and
torques of all phases are illustrated in Figure 8. Here, three phase PWM inverter operation
can be divided into six modes according to the current conduction states.

Fuzzy Controllers Recent Advances in Theory and Applications 72


Figure 7. Configuration of the BLDC motor and inverter system

Figure 8. (a) Active switches, (b) Back emf and phase current waveforms, and (c) Three phase torques
of the BLDC motor drive system
Analysis of the BLDC motor is based on the following assumptions for simplicity and
accuracy [3,10,15,16].
1. The BLDC motor is not saturated.
2. Stator resistances of all the windings are equal, self and mutual inductances are
constant.
3. Semiconductor devices in the inverter are ideal.
4. Iron losses are negligible.

Design of a Real Coded GA Based Fuzzy Controller for Speed Control of a Brushless DC Motor 73
Back-emf waveforms of all phases are equal under above assumptions. A BLDC motor
model can be represented as;

1 0 0 0 0
0 1 0 0 0
0 0 1 0 0
1 0 0
0 1 0
0 0 1
,
a a a
b b b
c c c
a
b s m
c
i / L v R / L i
d
i / L v R / L i
dt
i / L v R / L i
/ L e
/ L e ( L L - L )
/ L e
( ( ( ( (
( ( ( ( (
=
( ( ( ( (
( ( ( ( (

( (
( (
=
( (
( (

(8)
where va, vb, and vc are the phase voltages, ia, ib, and ic are the phase currents, R and Ls are the
stator resistance and inductance, Lm is the mutual inductance, ea, eb, and ec are the trapezoidal
back-emfs.
The motion equation is expressed as;

1
e
d
(T T B)
dt J
=

(9)

d

dt
=
(10)
where T is the load torque in Nm, J is the moment of inertia in kgm
2
, B is the frictional
coefficient in Nms/rad, and is the rotor speed in electrical rad/sec. The output torque is
redefined by Eq.(11) using back-emfs.


e a a b b c c
T (e i e i e i ) / = + +
(11)
This torque expression causes a computational difficulty at zero speed as the induced emf is
zero. In this study, the trapezoidal back-emf waveforms are modeled as a function of rotor
position to be able to estimate position actively according to the operation speed.
3.2. Modeling of trapezoidal back-emf
The back-emfs can be expressed as a function of rotor position () [17];


a a
b b e
c c
e f ()
e E f () , (E k )
e f ()
( (
( (
= =
( (
( (

(12)
where ke is back-emf constant, fa (), fb (), and fc () are the functions of rotor position as
shown in Figure 8. In this study, fa () trapezoidal function with limit values between +1 and
-1 expressed by Eq.(13) has been employed.

Fuzzy Controllers Recent Advances in Theory and Applications 74

6 0 6
1 6 5 6
6 6 5 6 7 6
1 7 6 11 6
6 12 11 6 2
a
( / ) ( / )
(/ / )
f () ( / ) ( / / )
( / / )
( / ) ( / )
< s

< s

= + < s

< s

< s

(13)
fb () and fc () can be determined in a similar way considering Figure 8. Substituting
equations (12)-(13) into equation (11), the output torque expression becomes,



e e a a b b c c
T k ( f () i f () i f () i ) = + +
(14)
3.3. Advanced simulation model of BLDC motor by using Runge-Kutta
numerical integration method
Runge-Kutta method is a frequently used method for solving differential equations
numerically. In engineering solutions, fourth order Runge-Kutta method is the most widely
used one [18]. For BLDC motor simulations, ia, ib, ic, and parameters, which have been
given by Eqs. (8), (9) and (10) are calculated by using fourth order Runge-Kutta method. For
example, the current associated with the a-phase (ia) is calculated using Eq. (15). Other
parameters are calculated in the similar way.

1
1
2
2
3
4 3
1 2 3 4
1
1
2
1
2
1
1 2 2
6





a a a
a a a
a a a
a a a
a a
k (v R i (kT) e )
L
k
k (v R ( i (kT) ) e )
L
k
k (v R ( i (kT) ) e )
L
k (v R ( i (kT) k ) e )
L
T
i ((k )T) i (kT) (k k k k )
=
= +
= +
= +
+ = + + + +
(15)
where k is the sample and T is the sampling period.
3.4. Reference current generator
Reference current generator determines reference phase currents (
* * *
a b c
i ,i ,i ) of the motor in
regard to reference current amplitude (I
*
), which is calculated using rotor position ().
Reference current amplitude (I
*
) can be obtained from Eq.(16).

*
t
I u / k =
(16)
where u is the control signal and kt is the torque constant. Phase currents given in Table 2
can be attained from Figure 8. These currents are input to PWM current control block [19].

Design of a Real Coded GA Based Fuzzy Controller for Speed Control of a Brushless DC Motor 75
Rotor position (-Degree)
Reference currents (A)
*
a
i
*
b
i
*
c
i
0-30 0 -I
*
I
*

30-90 I
*
-I
*
0
90-150 I
*
0 -I
*

150-210 0 I
*
-I
*

210-270 -I
*
I
*
0
270-330 -I
*
0 I
*

330-360 0 -I
*
I
*

Table 2. Reference currents of the BLDC motor
3.5. Current control block
In PWM current control block, reference phase currents (
* * *
a b c
i ,i ,i ) acquired from reference
current generator is compared with actual phase currents of the motor (
a b c
i ,i ,i ). These
current error values (
ia ib ic
e ,e ,e ) obtained using Eq. (17) are applied to inverter hysteresis
band (hb) and in regard to the switching states shown in Figure 8 (a), switching signals of
three-phase PWM inverter are generated [15,16].

*
ia a a
*
ib b b
*
ic c c
e i i
e i i
e i i
=
=
=
(17)
Inverter phase voltages (vao, vbo, vco) in reference to midpoint of DC supply voltage (Vdc) are
obtained using Eq. (18) by using current error values.

2
2
2
2
2
2
dc ia b
ao
dc ia b
dc ib b
bo
dc ib b
dc ic b
co
dc ic b
V / e h
v
V / e h
V / e h
v
V / e h
V / e h
v
V / e h
>
=

s

>
=

s

>
=

s

(18)
BLDC motor phase voltages (va, vb, vc) are given in Eq. (19) related to inverter phase voltages
determined from Eq. (18).

1
2
3
1
2
3
1
2
3
a ao bo co
b bo ao co
c co ao bo
v ( v v v )
v ( v v v )
v ( v v v )
=
=
=
(19)

Fuzzy Controllers Recent Advances in Theory and Applications 76
4. Simulation results
An algorithm has been developed to simulate the proposed real coded GA based fuzzy
controller in BLDC motor drive. In all simulations and practical applications, the BLDC
motor and inverter having the parameters listed in Table 3 have been used.

BLDC motor type Ametek 119003-01
Rating (P) 106 watt
Number of Phase (Connection) 3 (Star)
Rated speed 4228 rpm.
Rated current 6.8 A
Stator equivalent resistance (R) 0.348
Stator equivalent inductance (L) 0.314 mH
Moment of inertia (J) 0.0019 Ncm-s
2
Number of Pole (p) 8
Voltage constant (ke) 0.0419 V/rad/s
Torque constant (kt) 4.19 Ncm/A
PWM frequency (fPWM) 20 kHz
Inverter DC supply (Vdc) 24 volt
Inverter hysteresis limits (hb) 0.5
Inverter current limiter (Ibase) 20 A
Table 3. Parameters of the BLDC motor and Inverter
First of all, conventional fuzzy controller is designed based on trial-error method and
simulation. The parameters of conventional fuzzy controller are obtained as; n1=10, n2=1,
n3=2.2, a1=0.33, a2=0.66, b1=0.33, b2=0.66, c1=0.33, c2=0.66, R1=NB, R2=NB, R3=NM, R4=NM,
R5=NS, R6=NS, R7=Z, R8=PS, R9=PS, R10=PM, R11=PM, R12=PB, R13=PB.
During simulation of sequential approach, normalization parameters, then rule base and the
membership functions have been optimized sequentially. In this approach, the chosen
values for mutation probability is 0.05, crossover probability is 0.8, population number is 20,
the reference speed for the BLDC motor is set at 2000 rpm, and the motor operates in full
load. In the second approach which uses simultaneous optimization, the process is
implemented by coding all fuzzy controller parameters in GA simultaneously. In this
approach, the chosen values for mutation probability is 0.1, crossover probability is 1,
population number is 20, the reference speed for the BLDC motor is set at 2000 rpm and the
motor operates in full load. The obtained design parameters of optimal fuzzy controllers
using sequential and simultaneous real coded GA are given in Table 4.
For a reference speed of 2000 rpm and operating the BLDC motor in full load, the speed
responses and error variations of conventional fuzzy control, sequential GA based fuzzy
control and simultaneous GA based fuzzy control are shown in Figure 9 (a) and 9 (b),
respectively.

Design of a Real Coded GA Based Fuzzy Controller for Speed Control of a Brushless DC Motor 77
Fuzzy Controller Sequential Real Coded GA Simultaneous Real Coded GA
Norm. Parameters n1=642.348, n2=29.895, n3=3.144 n1=369.102, n2=8.448, n3=9.849
Parameters of
Membership Func.
a1=0.8488, b1=1.1002, c1=1.1309
a2=1.6005, b2=2.1, c2=1.5
a1=1.3035, b1=1.0762, c1=0.2051
a2=1.5191, b2=2.9, c2=1.62
Rule Base
R1=NM, R2=PM, R3=NB, R4=NM,
R5=Z, R6=NM, R7=Z, R8=PM, R9=Z,
R10=PM, R11=PB, R12=NM, R13 =PM
R1=NB, R2=PM, R3=NS, R4=NB,
R5=NB, R6=NM, R7=NS, R8=PM,
R9=PB, R10=PB, R11=PS, R12=NM,
R13=PB
Performance Index Jin= 3.2349 (For 150 generations) Jin= 3.2327 (For 850 generations)
Table 4. Optimal design parameters of Real Coded GA based fuzzy controllers

Figure 9. (a) Speed responses and, (b) Errors of conventional and GA based fuzzy controllers
The convergence of performance index (Jin) for sequential and simultaneous real coded GA
based fuzzy control system are shown in Figure 10 (a) and 10 (b), respectively.

Figure 10. Performance index of the (a) sequential, (b) simultaneous real coded GA based fuzzy control
system
In the sequential optimization method, the performance index (Jin) of the control system
reduces to 3.2349 and this process completes in 150 generations using real coded GA. In the
simultaneous optimization method, the performance index of the system (Jin) has been reduced
to 3.2327. This is the best optimization value obtained in this study. The optimization process

Fuzzy Controllers Recent Advances in Theory and Applications 78
with simultaneous GA completes in 850 generations. Also, the performance index value
indicates that simultaneous optimization process gives better results than the sequential
optimization process. However, while the overall controller is optimized in 150 generations in
sequential optimization method, the optimization is completed in 850 generations in
simultaneous approach, in which all parameters are optimized simultaneously.
5. Experimental results
Block diagram of the configuration of DSP based experimental system is shown in Figure 11.
The experimental system consist of a brushless DC motor, a voltage source inverter, a
current detector for hysteresis current control loop, back-emf detector for sensorless speed
control loop, ADSP-21992 Ez-Kit Lite evaluation board, interface devices between ADSP and
driver board, a PC and VisualDSP++ software for emulating and programming the DSP.

Figure 11. (a) Block diagram, (b) an overview picture of experimental setup of BLDC motor control
systems

Design of a Real Coded GA Based Fuzzy Controller for Speed Control of a Brushless DC Motor 79
In the experimental study, sensorless control of the BLDC motor was implemented
successfully using conventional fuzzy controller, sequential and simultaneous real coded
GA based fuzzy controller, which have parameters obtained by simulation and given in
Table 4. Optimal fuzzy controller was implemented in high level C programming language;
the program was compiled by VisualDSP++ C compiler and downloaded to the ADSP-21992
DSP controller board. The block diagram of the configuration of closed loop speed control of
BLDC motor executed without position and speed sensors is illustrated in Fig 11(a). The
experimental setup implemented in the laboratory is shown in Fig 11(b). The reference
speed for BLDC motor is set at 2000 rpm, and the motor operates in full load similar with
the simulations.
In Figure 12, phase to phase voltage and phase current waveforms of BLDC motor
are shown. The responses of both conventional and sequential real coded GA based fuzzy
controller are shown in Figure 13, conventional fuzzy and simultaneous real coded
GA based fuzzy controller are shown in Figure 14, sequential real coded GA based fuzzy
and simultaneous real coded GA based fuzzy controller are shown in Figure 15,
respectively.
The experimental results illustrated in these figures prove that the BLDC motor control
system was implemented successfully and it operates stably. Also, it is indicated that
simulation results show agreement with experimental results.




Figure 12. Phase to phase voltage and phase current of the BLDC motor. Time/Div:2.50msec,
Volt/Div:20.0V(CH1), Amper/Div:2.00A(CH2)

Fuzzy Controllers Recent Advances in Theory and Applications 80




Figure 13. Speed responses of sequential and conventional fuzzy control of the BLDC motor.
Time/Div:250 msec, Volt/Div:400 rpm.




Figure 14. Speed responses of simultaneous and conventional fuzzy control of the BLDC motor.
Time/Div:2.50 msec, Volt/Div:400 rpm.

Design of a Real Coded GA Based Fuzzy Controller for Speed Control of a Brushless DC Motor 81




Figure 15. Speed responses of sequential and simultaneous Real Coded GA based fuzzy controller.
Time/Div:250 msec, Volt/Div:400 rpm.
6. Limitations of the proposed methods
Although real coded genetic algorithm based fuzzy controller has shown a good behaviour
for sensorless speed control of a Brushless DC Motor, we think that there are some
limitations and disadvantages as follows in proposed method.
- Like other artificial intelligence techniques, the genetic algorithm cannot assure
constant optimisation response times. Even more, the difference between the shortest
and the longest optimisation response time is much larger than with conventional
gradient methods. This unfortunate genetic algorithm property limits the genetic
algorithms' use in real time applications.
- Genetic algorithm applications in controls which are performed in real time are limited
because of random solutions and convergence, in other words this means that the entire
population is improving, but this could not be said for an individual within this
population. Therefore, it is unreasonable to use genetic algorithms for on-line controls
in real time systems without testing them first on a simulation model.
- For the design of an effective controller, simulation model of the system is necessary to
obtain the most accurately. However, even if the best controller have been designed in

Fuzzy Controllers Recent Advances in Theory and Applications 82
simulation environment, fine tuning may be require in order to use the same controller
in on-line system.
- Micro genetic algorithm which can produce near optimal solution may be used in on-
line system for fine tuning.
- It is not possible to show the stability of the fuzzy controlled system, since the model is
not known. If the system model is obtained exactly, the use of tradional gradient
descent methods may give better result for optimal control in on-line systems.
- In fuzzy control systems, computing time could be long, because of the complex
operations such as fuzzifications and particularly defuzzification. So, the fuzzy
controller hardware requires high-speed processors, such as DSP, FPGA.
- The advantage of sensorless BLDC motor control is that the sensing part can be omitted,
and thus overall costs can be considerably reduced. In addition, the disadvantages of
sensorless control are higher requirements for control algorithms and more complicated
electronics.
7. Conclusions
As a result of the study, optimal fuzzy controller has been designed off-line by using real
coded GA and the obtained fuzzy controller has been used on-line for DSP-based BLDC
motor control system. Modeling of the BLDC motor was performed more accurately to
take account of trapezoidal back-emf waveforms and furthermore, fourth order Runge-
Kutta numerical integration method was used to decrease the truncation error and
numerical instabilities in simulation. Also, the results of our research indicate that an
improvement in the transient state and steady state responses of the system has been
obtained by means of optimization process using real coded GA. It is clear that, sequential
optimization takes less time. Also, observing the system speed response and error curves,
it can be deduced that the sequential method gives satisfactory results and that it can be
preferred in applications. It was observed that the use of real coded GA makes it possible
to adjust system parameters more precisely. Also, the size of chromosomes, in which a
great deal of parameters is coded, is reduced. Besides, the proposed method limits the
process time to minutes.
Author details
Omer Aydogdu and Ramazan Akkaya

Department of Electrical and Electronics Engineering,
Faculty of Engineering and Architecture, Selcuk University, Konya, Turkey
Acknowledgement
This study was supported by Selcuk University Scientific Research Projects (BAP) Support
Fund under contract number 2003/051.

Design of a Real Coded GA Based Fuzzy Controller for Speed Control of a Brushless DC Motor 83
8. References
[1] F. Cupetino, V. Giordano, D. Naso, B. Turchiano, L. Salvatore (2003) On-Line Genetic
Design of Fuzzy Controllers for DC Drives with Variable Load. IEE Electronics Letters,
Vol. 39 (5), pp. 479-480.
[2] G. Acosta, E. Todorovich (2003) Genetic Algorithm and Fuzzy Control: A Practical
Synergism for Industrial Applications. Elsevier Science Direct, Computer in Industry,
Vol. 52, pp. 183-195.
[3] O. Aydogdu (2006) Sensorless Control of Brushless DC Machines by means of Genetic
Based Fuzzy Controller. PhD. Thesis, Selcuk University, Turkey.
[4] F. Ashrafzadeh, E.P. Nowicki, M. Mohammadian, J.S. Salmon (1996) An Effective
Approach for Optimal Design of Fuzzy Controllers. IEEE Canadian Conference on
Electrical and Computer Engineering, pp. 542-545, Alberta, Canada.
[5] C. Xia, P. Guo, T. Shi, M. Wang (2004) Speed control of brushless dc motor using genetic
algorithm based fuzzy controller. Proc. of Int. Conf. on Intelligent Mechatronics and
Automation, pp. 460-464, Chengdu, Chine.
[6] H.P. Stpathy (2003) Real Coded for Parameters Optimization in Short-Term Load
Forecasting. Springer-Verlag Berlin Heidelberg IWANN, pp. 417-124.
[7] N. Hemati, M.C. Leu (1992) A complete model characterization of brushless dc motors.
IEEE Transactions on Industry Applications, Vol. 28 (1), pp. 172-180.
[8] O. Aydogdu, R. Akkaya (2005) DSP Based Fuzzy Control of a Brushless DC Motor
Without Position and Speed Sensors. Proceedings of 4
th
International Advanced
Technologies Symposium, pp. 182-187, Konya, Turkey.
[9] C.K Lee, W.H. Pang (1994) A Brushless DC Motor Speed Control System Using Fuzzy
Rules. IEE Power Electronics and Variable Speed Drives, pp. 101-106.
[10] V. Donescu, D.O. Neacsu, G. Griva, F. Profumo (1996) A Systematic Design Method for
Fuzzy Controller for Brushless DC Motor Drives. Proc. of the 27th. IEEE Annual Power
Electronics Specialists Conference, pp. 689-694, Baveno, taly.
[11] M. unka, R. Akkaya (2002) Compare with Binary and Real Coded Genetic
Algorithms. Selcuk University The Journal of Engineering, Vol. 7 (2), pp. 11-17.
[12] C.J. Wu, G.Y. Liu (200) A Genetic Approach for Simultaneous Design of Membership
Functions and Fuzzy Control Rules. Kluwer Academic Pub. Journal of Intelligent and
Robotic Systems, Vol. 28, pp. 195-211.
[13] N.J. Radcliff (1991) Formal Analysis and Random Respectful Recombination. In Proc. of
Fourth International Conference on Genetic Algorithms, San Diego, CA, USA.
[14] R.L. Haupt, S. Haupt (1998) Practical Genetic Algorithms, A Willey-Interscience
Publication, USA.
[15] B. Lee, M. Ehsani (2003) Advanced Simulation Model for Brushless DC Motor Drives.
Taylor & Francis Inc. Electric Power Component and Systems, Vol. 31, pp. 841-868.
[16] H.A. Toliyat, T. Gopalarathnam (2002) AC Machines Controlled as DC Machines
(Brushless DC Machines/Electronics). In: L.S. Timothyl, (eds) The Power Electronic
Handbook. CRC Press LLC, New York, pp. 78-100.

Fuzzy Controllers Recent Advances in Theory and Applications 84
[17] B. Sing, K. Jain (2003) Implementation of DSP based digital speed controller for
permanent magnet brushless dc motor. IE(I) Journal-EL., Vol. 84, pp. 16-21.
[18] J.R. Rice (1983) Numerical Methods, Software, and Analysis. McGraw-Hill, New York.
[19] T. Kim (2003) Sensorless Control of the BLDC Motors from Near-Zero to Full Speed.
PhD. Thesis, Texas A&M University, Texas, USA.
Chapter 4
A Type-2 Fuzzy Model Based on Three
Dimensional Membership Functions for
Smart Thresholding in Control Systems
M.H. Fazel Zarandi, Fereidoon Moghadas Nejad and H. Zakeri
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/47798
1. Introduction
This chapter focuses on the basic concepts of novel fuzzy sets, three dimensional (3D)
memberships and how they are applied in the design of type-1 and type-2 fuzzy
thresholding in control systems. Automatic fuzzification and membership functions shape
selection play a crucial role in fuzzy thresholding design and finally determination of
outputs via defuzzification. The related methodology and theoretical base will be discussed
in depth, using real examples in automatic control (Pavement distress detection and
classification). In spatial domain, selection of membership functions is a difficult task. It
should be noted that selection of a supper membership function is a golden key. This is one
of the major aims of this chapter to introduce a robust method to consider the uncertainty of
membership values by using flexible thresholding for controller problems.
In direct approach to fuzzy modeling, deep knowledge of expert plays a key role for
membership functions generation. In application, ambiguity of membership function
assignment is the main problem with fuzzy sets and systems. So, different fuzzy
membership functions may have various impacts on the systems and, then, different
thresholds in control problems.
To solve this problem, type II fuzzy thresholding is recommended. The upper and lower
membership functions promote this dilemma; however the figure of uncertainty (FOU) has
a fixed value that is equal to one, in upper and lower membership function. So, Type-2
fuzzy logic can effectively improve the control characteristics using FOU of the membership
functions.
In this chapter, a smart thresholding technique with its application will be presented, which
processes threshold as flexible type-2 fuzzy sets. The concept of ultra-fuzziness aims at

Fuzzy Controllers Recent Advances in Theory and Applications 86
capturing/eliminating the uncertainties within fuzzy systems using regular (type I) fuzzy
sets. A measure of 3D ultra-fuzziness is also presented. Several Experimental results are
provided in order to demonstrate the usefulness of the proposed approach.
We start with a real problem in control. The simplest method is to visually inspect the
pavements and evaluate them by subjective human experts. This approach, however, involves
high labor costs and produces unreliable and inconsistent results. Furthermore, it exposes the
inspectors to dangerous working conditions on highways. Destructive Testing (DT) and Non
Destructive Testing (NDT) are both costly and time consuming. To overcome the limitations of
the subjective visual evaluation process; several attempts have been made to develop an
automatic procedure (Moghadas Nejad and Zakeri, 2011,a,b,c) and (Daqi et al, 2009).
Most current systems use computer vision and image processing technologies to automate
the process. However, due to the irregularities of pavement surfaces, there has been a
limited success inaccurately detecting cracks and classifying crack types. In addition, most
systems require complex algorithm with high levels of computing power. While many
attempts have been made to automatically collect pavement crack data, better approaches
are needed to evaluate these automated crack measurement systems (Moghadas Nejad and
Zakeri, 2011,a,b,c) and (Daqi et al,2009)
A Hybrid Automatic Expert System (HAES) for automatic distress detection developed, based
on complex AI methods (Expert system, Polar Fuzzy Logic) and image processing methods
(Wavelet Transform, Inverse Wavelet Transform, 3D Radon Transform, Fast Fourier
transform, EH, etc). Fuzzy logic methods are one among favorite and overwhelming architect
that used for uncertainty simulations. Type-1 fuzzy sets (T1 FSs) have been successfully used
many area such as image processing, pattern recognition, machin learning. (Choi and Rhee,
2009), (Hagras,2004), (Hwang. Rhee, 2004), (Hwang. Rhee, 2007), (John, 2000), (Karnik, J.
Mendel, 1999), (Liang et al. 2000), (Liang, J. Mendel, 2001), (Makrehchi, et al. 2003), (Rhee,
2007), (Rhee, Choi, 2007), (Rhee, Hwang, 2001), (Rhee, Hwang, 2002) and (Rhee, Hwang, 2003).
Automatic generation of T1 FMFs classified as a interesting and hot research area. many T1
FMF generation models have been tested and various degree of successes achieved (Choi and
Rhee, 2009), (Makrehchi, et al.2003), (Medasani et al,1998), (Rhee, and Krishnapuram, 1993),
(Wang, 1994) and (Yang and Bose, 2006). Heuristics, histograms, probability, and entropy are
good tools to automate the T1 FMFs generation. Several methods under title of AI have been
implemented to data sets to generate T1 FMFs. A good classification proposed for T1 FMFs by
Choi and Rhee, (2009). Based on this classification, algorithms based on the fuzzy nearest
neighbor, back-propagation neural network, fuzzy C-means (FCM), robust agglomerative
Gaussian mixture decomposition (RAGMD), and self-organizing feature map (SOFM) were
used to generate T1 FMFs must be a considered as FMFs generator. (Choi and Rhee, 2009).
Uncertain meaning, uncertain measurement and noisy data are main causes that we cannot
obtain satisfactory results using T1 FSs, therefore in this mode employment of type-2 fuzzy
sets (T2 FSs) for managing uncertainty solved the problems (Ensafi & Tizhoosh, 2005), (Choi
and Rhee, 2009). Choi and Rhee (2009) stated that, because of the extra degree of freedom
(DOF), T2 FSs can control the blurring better than T1 FSs. However, undesirable amount of
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 87
computations stand in front of extension T2 FSs in vast scale applications. Interval type-2
fuzzy sets (IT2 FSs) are proposed to reduce the complexity (Choi and Rhee, 2009). Many
algorithms based on the T2 FMF have been proposed. (Choi and Rhee, 2009), (Hagras,2004),
(Hwang. Rhee, 2004), (Hwang. Rhee, 2007), (John, 2000), (Karnik, J. Mendel, 1999), (Liang et
al. 2000), (Liang, J. Mendel, 2001), (Makrehchi, et al. 2003), (Rhee, 2007), (Rhee, Choi, 2007),
(Rhee, Hwang, 2001), (Rhee, Hwang, 2002) and (Rhee, Hwang, 2003).
In this chapter, we focus on the generation of 3D Polar fuzzy Memberships functions to use
in hybrid expert system for systematic pavement distress detection and classification. In
particular, we consider 3D polar type-1 fuzzy membership functions (3D T1 PMFs) that are
generated from sample images and then developed to 3D polar type-2 fuzzy membership
functions (3D T2 PMFs). First, we review three methods based on heuristics, histograms,
and interval type-1 fuzzy C-means (IT1 FCM). For each method, the footprint of uncertainty
(FOU) is only required to be obtained, since the FOU can completely describe a T1 PMF. We
proposed two methods based on 3D domain and then 3D polar under the theory of type 2
fuzzy sets.
This paper is organized as follows.
In Section 2, we briefly review basic concepts and existing methods and background. In
Section 3, we managed the IT2 FMF generation methods. In Section 4, concepts of polar
fuzzy are discussed and we explain how our proposed IT2 PMF generation methods can be
implemented. Section 5 approximate reasoning and fuzzy inference discussed. Finally,
Section 6 gives the summary and conclusions.
2. Background
The extension of T1 FSs to T2 FSs can be used to effectively describe uncertainties in
situations where the available information is uncertain. T2 FSs consider as a blurred
membership function. The blurring used to model the uncertainty of crisp T1 FSs. A T2 FS
can be formulated as follow:



( )
( )
, 0,1
x
x X
A x
x
x
x X x X
f u
u
A XJ
x XJ

=
= e
e e
(
(

= = _ (

}
} }
(1)
where
( )
x
f u is the blurred membership function and
x
J is the original membership
(Mendel, 2001). Footprint of uncertainty (FOU) is a region between the blurred membership
function. The FOU of A
`
can be expressed by as

( ) { }
, : [0,1] by
x x A A x X
FOU A J x u u J FOU A and FOU A
= = =

| | | | | |
= = _ = = | | |
| | |
\ . \ . \ .


(2)
FOU constructed form upper membership function (UMF) and lower membership function
(LMF). (Choi and Rhee, 2009)

Fuzzy Controllers Recent Advances in Theory and Applications 88

Figure 1. A possible way to construct type II fuzzy sets. The interval between lower and upper
membership values (shaded region) should capture the footprint of uncertainty (FOU).
Although T2 FSs may be useful in modeling uncertainty, where T1 FSs cannot, the
operations of T2 FSs involve numerous embedded T2 FSs which consider all possible
combinations of secondary membership values. Therefore, undesirably large amount of
computations may be required. An effectively method to reduce the computational
complexity is interval type-2 fuzzy sets (IT2 FSs).
In General, FOU ( A
=
) can be expressed as: (Choi and Rhee, 2009)
FOU ,
x X
A FOU A FOU A
= = =

(
| | | | | |
= ( | | |
| | |
(
\ . \ . \ .

(3)
As a result, IT2 FSs requires only simple interval arithmetic for computing.
3. Automatic MF generators (AMFG)
In this section, we introduce a method for effectively crating IPT-1 FMF automatically from
images data. Several methods such as heuristics, histograms, and interval type-2 fuzzy C-
means (IT2 FCM) are proposed by (Choi & Rhee, 2009) for generating IT2 FMF automatically
from pattern data. Using scaling factor and heuristic T1 FMFs, IT2 FMF simply can be
generating. The histogram based method uses suitable parameterized functions chosen to
model the smoothed histogram for each class and feature extracted from sample data (Choi
and Rhee, 2009),(Hagras,2004), (Hwang and Rhee, 2004), (Hwang and Rhee, 2007), (John,
2000), (Karnik, J. Mendel, 1999), (Liang et al. 2000), (Liang, J. Mendel, 2001), (Makrehchi, et al.
2003), (Rhee, 2007), (Rhee and Choi, 2007), (Hwang and Rhee, 2001), (Rhee, Hwang, 2002) and
(Rhee and Hwang, 2003). The IT2 FCM based method uses the derived formulas of the IT2
FMFs in the IT2 FCM algorithm (Hwang and Rhee, 2002). A detailed description of each
method is discussed. The heuristic method simply uses an appropriate predefined T1 FMF
function, such as triangular, trapezoidal, Gaussian, S, or p function, to name a few, to initially
represent the distribution of the pattern data. The following are some frequently used heuristic
membership functions. (Choi and Rhee, 2009) Membership functions for fuzzy sets can be
constructed by any method exact, heuristic and Meta heuristic, such as triangular, trapezoidal,
Gaussian, S, or p function in the domain. Two most important constraints must be considered
for selecting a membership functions first, A membership function must be restricted between
[0 1] and the next A(x) must be unique. Four possible membership functions are presented in
Fig.2. Where type III and polar are new generation of fuzzy membership function that can be
used in several application in the control and classification domains. In the field of pavement
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 89
management system this new generation of MF play a powerful link between several tools
such as multi-resolution methods (wavelet and beyond the wavelet methods), image
processing, NN and expert system.

Figure 2. Possible membership functions
A possible membership function can be defined for every category by expert with any tools.
For example using image processing techniques and Radon transform, several membership
function generated and shows in Fig.3 for pavement cracking distress.

Figure 3. Variety of membership functions

Fuzzy Controllers Recent Advances in Theory and Applications 90
More simple and complex functions can be used under the form of discrete and continues.
Generally the ordinary functions categorized in Triangular, Trapezoidal, -membership, S-
membership, Logistic, Exponential-like and Gaussian function. Additionally several more
advanced membership function which generate by automatic generator introduced. More
applications in image processing frequently used heuristic membership functions that can
be generally categorized in Table.1.
- Triangular function

( ) ( )
( ) ( )
, 0
, /
, /
, 0
IF x a then
IF a x b Then x a b a
IFa x b Then c x c b
IF x c then

s =
s s =
s s =
> =

- Trapezoidal function

( ) ( )
( ) ( )
, 0
, /
, 1
, /
, 0
IF x a then
IF a x b Then x a b a
IFb x c Then
IFc x d Then c x d c
IF x d then

s =
s s =
s s =
s s =
> =

- Gaussian function
( )
2 2
( ) / 2
( )
x c
x e
o

=
- S-function
- b = (o + b)2
( ) ( ) ( )
( ) ( ) ( )
2
2
, 0
, 2 /
, 1 2 /
, 1
IF x a then
IFa x b Then x a b a
IF a x b Then x a c a
IF x c then

s =
s s =
s s =
> =

- P-function

( )
( )
, ( ; , , )
2
, ( ; , , )
2
b
IF x c then x s x c b c c
b
IF x c then x s x c b c c
t
t
s =
> =

Table 1. Heuristic membership functions, (Choi & Rhee,2009)
By control parameters, one can select a various interval pattern. Theses parameters usually
trained and learned by experts. Under the title of Control Parameter (o), the UMF of the IT2
FMF and LMF can be designed. The LMF and UMF determined by scaling o onJ [ between
0 and 1, which can be also tuned in supervised and unsupervised manner or provided by an
expert. Choi & Rhee, (2009) proposed a simple definition for FOU, which categorized in
heuristic methods. For feature i

FOU ,
,. . ,
x X
x X x X
A FOU A FOU A
FOU A FOU A or FOU A FOU A |
= = =

= = = =

(
| | | | | |
= ( | | |
| | |
(
\ . \ . \ .

( (
| | | | | | | |
= ( ( | | | |
| | | |
( (
\ . \ . \ . \ .

(4)
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 91
0 , 1 o | < <
to generalization, Choi & Rhee, (2009) choose the min operation as intersection for obtain the
overall FOU by taking intersections of all upper and lower memberships.
FOU ,
min , min . . , min
x X
x X
x X
A FOU A FOU A
FOU A FOU A or min FOU A FOU A |
= = =

= = = =

(
| | | | | |
= ( | | |
| | |
(
\ . \ . \ .

(
(

| | | | | | | |

(
( = | | | | ` ` ` `
| | | |
(
(
\ . \ . \ . \ .
) ) ) )
(

(5)
where FOU A
=
| |
|
|
\ .
and FOU A
=
| |
|
|
\ .
are the minimum UMF and LMF among all UMFs and
LMFs, respectively. Heuristic method which proposed Choi & Rhee (2009) is summarized in
Fig.4.
Histogram based method (HBM) for membership function generation is another method
which is more flexible than heuristic methods. In HBM, distribution of the feature values,
have a crucial role in T1 FMF determination elements. Choi & Rhee, (2009) clearly stated
that,membership functions generated from HBM may be considered more suitable for
arbitrary distributed data than from heuristics. Based on this theory Choi & Rhee (2009)
propose a new method for generation IT2 FMFs. Using smoothed histograms which
generated by hyper-cube or triangular window and then normalized, the upper and lower
membership function flourished and mapped to real data. Selection a well trained
parameters function to model the smoothed histograms has a tangible ramification on
performances of MF generator system. To avoid over fitting lowest, the suitable degree of
the polynomial function (PF) is stood out as the knee point of error. As a result, HBM FSs
requires good estimation of PF.
In our case, as a real example in control, Type, severity and extents of cracking in pavement
surface transform in a transform realm to generate a simple features. Simple features can use
for generation of T1 FMF. Approximate parameter values such as the number, height, and
location of peaks which related to cracking used to determine the optimal parameter values
of the function.
Choi & Rhee (2009) considered Gaussian functions as suitable to model the IT2 FMFs (Rhee
and Krishnapuram, 1993). They used a heuristic approach (Choi & Rhee, 2009) to obtain the
initial parameters. Choi & Rhee (2009) ignored the ones that have small peaks. This means
that we have a threshold that it considers as a crisp threshold. The FOU A
=
| |
|
|
\ .
and FOU A
=
| |
|
|
\ .

are obtained by again fitting PFs to the smoothed histograms. New again histograms
crystallized upper and lower MFs fitted to PF. As dimensional parameters or overall size of
problem increase, undesirably become more and more. These complexities arise due to the
high process in smoothing and fitting. This is a challenging point that set in motion to

Fuzzy Controllers Recent Advances in Theory and Applications 92

Figure 4. Heuristic based IT2 FMF generation method
product a new heuristic to handle computational load. Choi & Rhee (2009) proposed two
steep methods 1) calculate one-dimensional 1DUMF and 1DLMF for HBM. 2) Obtain the
overall FOU A
=
| |
|
|
\ .
and FOU A
=
| |
|
|
\ .
by Intersections operation. Intersections operation which
proposed for this aggregation expressed as

FOU , ,
i i
x X
x X
A FOU A FOU A min FOU A min FOU A
= = = = =

(
(
| | | | | | | | | |

(
= = ( | | | | | ` `
| | | | |
(
(
\ . \ . \ . \ . \ .
) )
(

(6)
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 93
where FOU A
=
| |
|
|
\ .
is the UMF, FOU A
=
| |
|
|
\ .
is the LMF, and i is the feature's number. From our
points of view, the main contributions of Choi & Rhees methods are developing in
memberships generation. These methods enable them to transfer the knowledge when
expert facing with N dimensional features. These methods are applicable for images realm.
We assert that this contribution is valuable. Nevertheless we would like to highlight that
high process in discrete smoothing and fitting(first 1DUMF and 1DLMF calculation and
then aggregation) faced us to problem to products an effective MF generator. Heuristic
method to generate T2 FMFs, which proposed Choi & Rhee (2009) is summarized in Fig.5.
Choi & Rhee (2009) considered fuzzy C-means (FCM) functions to model the IT2 FMFs
(Hwang and Rhee, 2007) (Choi & Rhee, 2009). The fuzzifier m in FCM, can be fired as a
membership generator. IT2 FCM based method proposed by Choi & Rhee (2009). They
stated that, Due to the constraint on the memberships we cannot design this region with any
particular single value of fuzzifier m to be used in the FCM. IT2 FCM algorithm was proposed
to solving this problem (Hwang and Rhee, 2007). Indeed they products a simple dynamic
fuzzifuyer AMFG to generating the Membership function. According to IT2 FCM, two
fuzzifier m1, m2 are employed to control the blurring area in fuzzy domain. The proposed
IT2 FMF in IT2 FCM expressed as (Choi & Rhee, 2009).

( )
, : [ , ]
x
J x u u FOU A FOU A
= =

| | | |

=
| | `
| |
\ . \ .
)

(7)

( ) ( )
( )
( )
1 2
1
2
2 2
1 1
1 1
2
1
1
2
1
1
1 1

1

1

m m
C C
ij ij
K K
ik ik
m
C
ij
K
ik
m
C
ij
K
ik
IF
d d
d d
FOU A THEN FOU A
d
d
ELSE FOU A
d
d

= =
= =

=
=

=




>


| | | |

| |
| |
\ . \ .

| | | |

= =
| | `
| |
\ . \ .
| |

|
|
\ .

| |
=
|
|

\ .
| |

|
\ .
)

, (8)
and

Fuzzy Controllers Recent Advances in Theory and Applications 94

( ) ( )
( )
( )
1 2
1
2
2 2
1 1
1 1
2
1
1
2
1
1
1 1

1

1

m m
C C
ij ij
K K
ik ik
m
C
ij
K
ik
m
C
ij
K
ik
IF
d d
d d
FOU A THEN FOU A
d
d
ELSE FOU A
d
d

= =
= =

=
=

=




s


| | | |

| |
| |
\ . \ .

| | | |

= =
| | `
| |
\ . \ .
| |

|
|
\ .

| |
=
|
|

\ .
| |

|
\ .
)

(9)
However IT2 FCM for updating cluster prototypes requires type-reduction. Using type-2
fuzzy operations therefore is essential. The crisp center obtained mean of centers of
defuzzification as the centroid obtained by the type-reduction according Eq.10

1 1
1
1
1
,
x x
FOU A J FOU A J
x
m
N
i
i
i
m
N
i
i
V C C
x FOU A
FOU A
= =
| | | |
e e | |
| |
\ . \ .
=
=
=
=
.
(
= =

| |
|
|
\ .
| |
|
|
\ .

(10)
The UMF and LMF for class k and input pattern xj can be expressed by modifying

{ }
{ }
*

( , ) ,
k
kj p j x k
d min d x V p n =


(11)
Based on Choi & Rhees (2009) method the membership values for the UMFs and LMFs are
based on m1 and m2 and they are highly dependent on value selection of threshold which is
itself considered crisp. Choi & Rhees (2009) stated that IT2 FCM can desirably control the
uncertainty that is quite simple handle all features of high dimensional problems. Their
heuristic method summarized in Fig 6.
The accuracy of IT2 FCM highly dependent on fuzzifiers selection. These parameters have
significant role in designing the FOU for a data set. In general, select unsuitable fuzzifier
worth poor clustering. (Choi & Rhees, 2009)
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 95

Figure 5. Heuristic based IT2 FMF generation method

Fuzzy Controllers Recent Advances in Theory and Applications 96

Figure 6. Heuristic based IT2 FCM generation method
4. Interval type-2 Polar Fuzzy Method (IT2 PFM)
4.1. Type III-MF
The Interval type-2 Polar Fuzzy Method (IT2 PFM) algorithm was proposed to
automatically control the uncertainty. In this section, we proposed an intelligent IT2 FMF
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 97
generator agent. First, the IT2 FMF algorithm introduced, and then our IT2 FPM based
method are described. We selected Cubic Smoothing Spline (CSS) for generate the upper
and lower membership functions because of non-uniform illumination of the Three
Dimensional Memebership Functions (3DMFs). In the Type-2 domain, the estimation of the
3DMFu and 3DMFL are exanimate from the fitting of a cubic smoothing Spline,( Mora et
al.,2011) to the 3DMF(x,y). The select CSS is a special class of Spline that can capture the low
3DMF value that limited the non-uniformity of the 3DMF (Culpin, 1986). The fitting
objective is to minimize the equation.
( ) ( ) ( )
2 2 2
1 1
. ( , , ) (1 ) ( , ) ,
m n
y x
M P f x y s x y p D S x y dxdy
= =
= +

(12)
where, this equation include two parts:
- Compactness: measures how close the spline is to the data that reflect to the summation
term which weighed by the smoothing factor p,
- Smoothness: measures the spline smoothness using its second derivative that reflect to
the integral term weighed by (1 - p).
The smoothing factor p, controls the balance between being an interpolating spline crossing
all data points (with p = 1) and being a strictly smooth Spline (with p = 0). The smoothing
spline f minimizes when
( ) ( ) ( ) ( )
2 2
2
1
:, ( ) (1 ) ( ) .
n
j
P j y j f x j p t D f t dt e
=
(
(
( | |
(
(
+
| (
(
( \ .


}
(13)
where, |z|2 represent for the sum of the squares of all the entries of n, N and M is the
number of entries of x and y, and the integral is over the smallest interval containing all the
entries of x and y. The default value for the weight vector w in the error measure is ones
(size(x)). The default value for the piecewise constant weight function in the roughness
measure is the constant function 1. Further, D2f denotes the second derivative of the function
f. The default value for the smoothing parameter, p, is chosen in dependence on the given
data sites x and y (Pal and Bezdek, 1994). The smoothing parameter determines the relative
weight to place on the contradictory demands of having f be smooth vs having f be close to
the data. For p = 0, f is the least-squares straight line fit to the data, while, at the other
extreme, i.e., for p = 1, f is the variational, or natural cubic spline interpolant. As p moves
from 0 to 1, the smoothing spline changes from one extreme to the other. (See Fig. 7)
The interesting range for p is often near
( ) ( )
3
1 / 1 , / 600 , min N M
| |
|
\ .
+ with h the average
spacing of the data sites, and it is in this range that the default value for p is chosen. For
uniformly spaced data, one would expect a close following of the data for p = 1/(1 +
(min(N,M))
3
/6000) and some satisfactory smoothing for

Fuzzy Controllers Recent Advances in Theory and Applications 98

Figure 7. As p moves from 0 to 1, the smoothing spline changes from one extreme to the other.

( ) ( )
3
1 / 1 , / 60 . 1 p min N M p
| |
|
\ .
= + > (14)
can be input, but this leads to a smoothing spline even rougher than the variational cubic
spline interpolate (Pal and Bezdek,1994).

3 3 3
1 1 1
, ,
( ( , )) ( ( , )) ( ( , ))
1 1 1
600 600 600
U L
U
L
p p p
min N M min N M min N M
|
| |
| | | |
|
| |
|
| |
= = =
|
| |
| + + +
| |
|
\ . \ .
\ .
(15)
A reference smoothing factor
( )
1 4 p e = was obtained empirically for constructed
MF in upper bound and
( )
0.93 5 p e = for constructed MF in lower bound. For example,
in the case of image thresholding, After testing several thresholds, the general rule can be
extract from 3DRT thresholds for upper and lower bounds by good selection o and [.
4.1.1. A measure of ultrafuzziness
Using a simple method, we turned ultrafuzzy to the 3DRT fuzzy set. According a type II
membership function, MF must be in [0,1]. One can be taking out the normalization form
3DMF using division every point by max 3DRT.
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 99

1
(3 )( , )
( , )
(3 )( , )


,

h
h
DRT i j
i j
h
DRT i j
v
RTMF
max v
| |
(
|

=
|
(
|

\ .
(16)

1
(3 )( , )
( , )
(3 )( , )


,

h
h
DRT i j
i j
h
DRT i j
v
H
max v

| |
(
|

= +
|
(
|

\ .
(17)
and

( )
1
(3 )( , )
( , )
1 1
(3 )( , )


1
,

h
h
N M
DRT i j
i j
h
h
j i
DRT i j
v
GR
MN max v = =
| |
(
|

=
|
(
|

\ .

(18)

where, M and N denotes the size of 3DMF platform,H is high platform, (1, ) h and
(3 )( , )
h
DRT i j
v is 3DRT value in the position i and j. Select a bigger h is worth a more enhanced
distress for example in pavement distress detection and classification problem and smoother
noisy background (see Fig.8). In order to define a type II fuzzy set, one can define a type I
fuzzy set and assign upper and lower membership degrees to each element to (re)construct
the footprint of uncertainty (Fig. 9) (Tizhoosh,2005). For example, when Radon Transform is
applied to wavelet modulus, a distress (crack) is transformed into a peak in radon domain.
Originally, every distress reflects to RT and has different intensity in 3DMF histograms. For
example mean of 3DRT have variety range. According to the above Eq. 18 the max GR must
be equal 1. To extend the fuzzy membership to type II fuzzy sets, ultrafuzziness should be
zero, if the MF can be selected without any ambiguous such as type I. The amount of
ultrafuzziness will increase by rising uncertainly bound.
The extreme case of maximal ultrafuzziness, equal 1, is worth to completely vagueness. pal
and bezdek (1994) had extensive reviewed well known fuzziness index, two general classes
proposed by them was additive and multiplicative class (Pal and Bezdek,1994). Based on
kufmanns Index of fuzziness for a set
( )
,
n
A r x e

( )
2
( , )
near
ka
k
H A d A A
n
| |
=
|
\ .
(19)
Where, k R
+
e , d is a metric, and A
ncu
is the crisp set close to the A. generally, based on d,
weigh of k determined. The J(A, A
ncu
) and linear or quadratic ( )
ka
H A cab be determined
by q-norms,

Fuzzy Controllers Recent Advances in Theory and Applications 100

1

,
1
( , )
near
q n
q
near
i
A i
i
d A A
=
| |
| =
|
\ .

1

1 ,
1
2
( , )
near
q n
q
ka i
A i
i
q
H q A
n

=
| |
| |
|
| =
|
|
|
\ .
\ .

(20)
Where q |1, ). On the other side, Tizhoosh developed a simple ultrafuzziness index for
the special case as fallow (Tizhoosh, 2005),
( )
1 1
1 1
1
( ) ( )
M N
U ij L ij
i j
A g g
MN


= =
(
=

(21)
where
( ) ( )
1/
U A
g g
o

(

= and
( ) ( )
1/
U A
g g
o

(

= , (1, 2]. o e and in general term it present
as follow,

( )
1
0
1
( ) ( ) ( )
L
U ij L ij
g
A g g h g
MN

=
(
=

(22)
Based on these theory and with respect to Tizhooshs method (Tizhoosh, 2005),for developing
ultrafuzziness on 2D data, a measure of ultrafuzziness y for a platform 3DMF with M*N sets,
surf 3DMF and the membership function
( , ) i j A

can be developed as follows:
( )
( )
( ) ( )
1
1 1
, , 1
1 1
1

q
M N q
u i j L i j
j i
q
A
MN


= =
| |
(
|
(
=
|
(
|
|
\ .

( )
( )
( ) ( ) ( ) ( )
1
1 1
( , )
1 , , , ,
1 1
1
0
q q
M N
i j
u i j T L i j T
j i
q
RT
T T
MN


= =
| |
(
|
c c
(
= =
|
c c (
|
|
\ .

(23)
This basic definition relies on the assumption that the singletons sitting on the FOU are all
equal in height (which is the reason why the interval-based type II is used), (Tizhoosh,
2005). The variation in the space can be measured by this method, therefore the new Index
introduced in three dimensional domain of FOU for 3D fuzzy sets, (3DFOU). This method
can resolve the problems about the ultrafuzziness index -uncertainty (FOU) has a constant
value, that equals one, in all the intervals of the universe of discourse (Ioannis et al., 2008) -
using introducing flexible membership function across the intervals path (see Fig 9, 10).
Similarly, We are evaluated, proposed method, based four conditions Minimum
ultrafuzziness, Maximum ultrafuzziness, Equal ultrafuzziness and Reduced ultrafuzziness that
every measure of fuzziness should satisfy, which introduced by Kaufmann (Kaufmann,
1975). In a similar way, we established that the new index is qualified for measure of
ultrafuzziness in 3D domain with these conditions.
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 101

Figure 8. Basic rules for construction 3D fuzzy type II memebership function.

Figure 9. Three dimensional domain of FOU for 3D fuzzy sets, (3DFOU)

Figure 10. Example of Three FOU for 3D fuzzy sets, (3DFOU) using proposed algorithm

Fuzzy Controllers Recent Advances in Theory and Applications 102
1. IF
( )
, i j
consider as a type I fuzzy set Then
( ) ( )
, , u i j L i j
= AND
( )
0 A =

( )
( )
( ) ( )
1
1 1
, , 1
1 1
1
0 [1, ).
q
M N q
u i j U i j
j i
q
A for q
MN


= =
| |
(
|
(
= = e
|
(
|
|
\ .

(24)
2. IF
( ) ( )
, ,
1
u i j L i j
= (high ambiguity) Then
( )
1. A =

( )
( )
( ) ( )
1
1 1
, , 1
1 1
1
1 [1, ).
q
M N q
u i j U i j
j i
q
A for q
MN


= =
| |
(
|
(
= = e
|
(
|
|
\ .

(25)
3.
( )
`
( )
A A =

Where (A

)is type II fuzzy set and its complement set can be

determined by 1-
( )
, u i j
and 1-
( )
, L i j
, therefore complement set defined as follow
For the complement set, the ultrafuzziness y is equal:

`
( )
( )
( ) ( )
( ) )
1
1 1
, , 1
1 1
1
1 1 1,
q
M N q
u i j L i j
j i
q
A A for q
MN


= =
| |
(
|
( = = e
|

(
|
|
\ .



(26)
4. IF
( , ) ( , )
3 3
i j d c
DFOU DFOU < Then
( ) ( )
( , ) ( , ) i j d c
A A <


.
4.1.2. Finding the optimum interval 3DMF
The general approach for 3DMF based on upper and lower MF is equal:

( )
( )
( ) ( )
min ,
1 , , 1 , ( ) 1
max ,
L ij U ij
i j
SURF i j g OR g
i j

(
( ( ( = = +

(

(27)
Where
( )
is ultra fuzzy coefficient and
( )
, i j

is ultra fuzzy value for

( )
, u i j
and
( )
, L i j
, in
upper and lower threshold.
4.2. Interval type-2 polar based method
Image processing is one among interesting applications of 3DMF. Instead of type reduce
from Type-2 to type-1, we used a polar transform to make uniformity by same scale in
[0,2n]. The RT of a two-dimensional function
( )
, f x y in
( )
, r u plane is defined as:

( ) ( ) ( ) ( ) ( )
, , , , , P r R r f x y f x y r xcos ysin dxdy u u o u u
++

(
= =

} }
(28)
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 103
where
( )
, f x y represents an image,
( )
, P r u is the radon transform of
( )
, f x y , represents
the line direction, and r is the distance away from the origin of coordinates. (Radon, 1919),
(Miao et al., 2012) Where
( )
o is the Dirac function, re[,] is the perpendicular distance of a
line from the origin and e[0,] is the angle formed by the distance vector. For the spatial case
such as 3DMF, the fuzziness can be calculated as follows (Tizhoosh, 2005);

( )
( )
,
,
1 1
( )
1 1
2
( ) ,1
i j
i j
M N
l A r
A r
j i
A min
MN


= =
| |
(
(
|
(
=
(
|
(
|

\ .

, (29)
where M N is subset A X _ with L radon transform value, 0, 1 r L e (

, the histogram
h(RT) and the membership function X(RT), the linear index of fuzziness l can be defined as
follows (see Fig.6):

( )
( )
1
( )
0
2
( ) ( ) ,1
L
l A r A r
r
A h r min
MN

=
| |
(
| (
= (
| (

( |

\ .

, (30)
To quantify the object fuzziness, a suitable membership function
( )
A
r should be
determined. Tizhoosh present different functions, such as the standard S-function, the
Huang and Wang function, LR-type fuzzy number (Tizhoosh et al, 1998; Huang and Wang,
1995; Pal and Bezdek, 1994; Pal and Murthy, 1990). Similar 3DMF presented in section 4.1, to
generation of polar MF, CSS is used. The estimation of the MF also exanimate from the
fitting of a cubic smoothing Spline, (Mora et al.,2011) to the 3DPMF(r,0). The fitting objective
is to minimize the equation.
( ) ( ) ( )
2 2 2
1 1
. ( , , ) (1 ) ( , ) ,
m n
y x
M P f r s r p D S r drd u u u u
= =
= +

(31)
where, this equation include two parts: Compactness and Smoothness. The smoothing factor p,
controls the balance between being an interpolating spline crossing all data points (with p =
1) and being a strictly smooth Spline (with p = 0). In the polar transform, as p moves from 0
to 1, the smoothing spline changes from one extreme to the other. (See Fig. 11)
Using Radon transform for MF generation have several benefits such as Translation,
Rotation and Scaling in IT2 FPM. (Miao et al., 2012).

( ){ } ( )
0 0 0
, ( , , R f x x y y P r r u u = , (32)

( ){ } ( )
, ( , , R f xcos ysin xcos ysin P u u + + = + (33)

( )
, , , , 0
y
x r
R f P u u


| | | |
= =
` | |
\ . \ .
)
(34)

Fuzzy Controllers Recent Advances in Theory and Applications 104

Figure 11. A sample of polar memberships function, As p moves from 0 to 1, the smoothing spline
changes from one extreme to the other.
Where
0 0 0
r x cos y sin u = + , y is the scaling factor and is the rotation angle. A rotation of
( )
, f x y by angle leads to a translation of
( )
, P u in the variable . A scaling of
( )
, f x y
results in a scaling in the p coordinate, as well as an intensity scaling of
( )
, P u . (Miao et al.,
2012). For the Fuzzy Polar based Method, we proposed use the following heuristic
approach. This method consists of seven steps to obtain the 3D membership function in the
polar domain.
Step 1. Three Dimensional Surface (3D Data), Using Radon transform generate the 3D
surface from image and construct 3D data surface.
Step 2. Three Dimensional Polar Surface (3D Polar), Transfer data to the polar domain and
uniform data in multi-scale.
Step 3. Polar Histogram Generator (PHG), generates polar histogram in all direction using
polar histogram generator.
Step 4. Approximate Smoother fitting parameters (SF), Perform SF parameter to obtain the
approximate parameter value (p).
Step 5. Polar Smooth Generator (PSG) smooths the histogram of the overall polar surface.
Step 6. Perform PSG fitting for the upper and lower histogram values.
Step 7. Determine PFMF, by normalizing the height of the upper PSG and LMF by the
lower PSG.
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 105
On advantages of T2 PFM method is decrease on computational load in comparison with
histogram based IT2 FMF. According our proposed method computational load can
decrease, due to the stimulatory dimension in muli-scale surface and decrease
computational load because of modified histogram smoothing process and fitting. Instead
finding the one-dimensional UMF and LMF for each class label and feature which used by
histogram based method, we fired all points in polar system with a cubic-spline. Next, we
obtain the overall UMF and LMF Simultaneously. To obtain the generated IT2 PMF, it
essential the three polar FOU (3D PFOU) be calculated. The UMF and LMF are designed by
refitting cubic-spline. According proposed method the smoothed histograms have values
that are above or below the mother fitted surface. Fig. 12a shows the one example
constructed by polar upper and lower cubic-spline functions.




Figure 12. A)Three dimensional domain polar upper bound and lower bound, b) FOU for 3D polar
fuzzy sets, (3D PFOU)
Using the upper and lower mother surface, to obtain the 3D PFOU, the PUMFs and PLMFs
surface are designed. The UMF and LMFs surface normalized. Fig. 12,b. shows the IT2 PMF
obtained by our proposed method. The shaded region between the 3D UMF and 3D LMF
indicates the 3D PFOU. As shown in the Fig.13, our proposed method can effectively design
IT2 FMFs based on the distribution of the input 3D data. The 3D PFOU can be expressed as
PFOU , min , min
x X
x X
A PFOU A PFOU A PFOU A PFOU A
= = = = =

(
(
| | | | | | | | | |

(
= = ( | | | | | ` `
| | | | |
(
(
\ . \ . \ . \ . \ .
) )
(

(35)
where PFOU A
=
| |
|
|
\ .
is the UMF, PFOU A
=
| |
|
|
\ .
s the LMF, and i is the feature number.

Fuzzy Controllers Recent Advances in Theory and Applications 106


Figure 13. Heuristic based IT2 FPM generation method
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 107
4.2.1. A measure of polar ultrafuzziness
Polar ultrafuzzy can be calculated based on 3DMF fuzzy set. Such as defuzzifcation method
proposed in measure of surface ultrafuzziness in section 4.1.1., type II membership function
must be in [0,1]. Similaraway, normalization must be used for 3D PMF generation by
division every point at ( , ) u by max 3D PRT.


1
(3 )( , )
( , )
(3 )( , )


,

h
h
D PMF
h
D PMF
v
PN
max v
u
u
u
| |
(
|

=
|
(
|

\ .
(36)

1
(3 )( , )
( , )
(3 )( , )



h
h
D MF
h
DRT
v
H
max v
u
u
u

| |
(
|

= +
|
(
|

\ .
(37)
and

( )
1
2
(3 )( , )
( , )
0 0
(3 )( , )


1
,
2 2
h
h
r
DMF
i j
h
h
DRT
v
GR d d
r max v
| |
(
|

=
|
(
|

\ .
} }
t
u
u
u
t
(38)
where, 2nr denotes the size of 3DPMF platform,H is high polar platform [0,1] H , (1, ) h
and
(3 )( , )
h
DRT
v
u
is 3DMF value in the position p anu 0. In thresholding, selection H
controller can use for select an optimum threshold based on type II fuzzy. Select a bigger H
is worth a more enhanced maximum value. In order to define a type II fuzzy set in polar
domain, first we develop a type I fuzzy set and assign upper and lower membership
degrees to each element to (re)construct the footprint of uncertainty in polar system (Fig.
14). Hear we select H=0 to calculate the real 3D PMF. In polar system the definition for
uncertainty is slightly deferent 3D FMF. Uncertainty can present in ring and height which
reflect to polar memberships function. (Fig. 16) For example, when Radon Transform is
applied to wavelet modulus, a distress (crack) is transformed into a peak in radon domain.
Originally, every distress reflects to RT and has different intensity in 3DMF histograms. For
example mean of 3D PRT have variety range. According to the above Eq. 38 the max GR
must be equal 1. Similaty 3D FMF method, in 3D PMF, the amount of ultrafuzziness will
increase by rising uncertainly bound.
The extreme case of maximal ultrafuzziness in polar system, equal 1, is worth to completely
vagueness. Based on Pal and Bezdek (1994) research on several fuzziness index, two general
classes proposed by them was additive and multiplicative class (Pal and Bezdek, 1994).
Based on Kufmanns Index of fuzziness for a set P r
n
(x),

Fuzzy Controllers Recent Advances in Theory and Applications 108

( )
2
( , )
near
ka
k
H A d P P
n
| |
=
|
\ .
(39)
Where, k R
+
e ,d is a metric, and
near
A is the crisp set close to the P. generally, based on
d,weigh of k determined. The ( , )
near
d P P and linear or quadratic ( )
ka
H P can be determined
by q-norms such as 3D FMF,

1 1

( , )
,( , ) 1 ,
1 1
2
( , ) , , ) (
near near
q q n n
q q
near
ka i
A P i
i i
q
d P P H q P
n
u
u

= =
| |
| | | |
|
| | = =
|
| |
|
\ . \ .
\ .

(40)
Where
)
1, q e

. Based on Tizhoosh ultrafuzziness index, we developed a new index in

continues polar domain (Tizhoosh, 2005),

( ) ( )
2

( , ) ( , )
2
0 0
1
( ) ( ) ,
4
r
U L
P g g d d
t
u u
u
t
=
} }

(41)
Where
( ) ( )
1/
U A
g g
o

(

= and
( ) ( )
1/
,
A L
g g
|

(

= and in general term it present as
follow,

( ) ( )
2 1

( , ) ( , )
2
0 0 0
1
( ) ( ) ( ) . .
4
r
U L
P g g h g d d dh
t
u u
u
t
=
} }}

(42)
A measure of ultrafuzziness y for a polar 3D FMF in
2
4t , polar 3D FMF and the
membership function
( , ) P u

can be developed as follows:

( )
( )
1
2
( , ) ( , )
1
0 0
1
( ) ( ) ,
q r
q
U L
q
P g g
Area
t
u u

| |
(
|
( =
|
(
|

|
\ .
} }

( )
( )
1
2

( , ) ( , ) ( , )
1
0 0
1
( ) ( ) 0
q r
q
U L
q
PMF g g d d
T T
Area
t
u u u
u
| |
(
|
c c
( = =
|
c c
(
|

|
\ .
} }

(43)
The variation in the polar space can be measured by this method, therefore the new Index
introduced in polar dimensional domain of FOU for 3D polar fuzzy sets, (3D PFOU). This
method can resolve the problems about the discontinues domain and in a same time reduce
on dimension by using polar transform. (see Fig 15, 16).
Similarly, Kaufmann conditions consists of Minimum ultrafuzziness, Maximum ultrafuzziness,
Equal ultrafuzziness and Reduced ultrafuzziness evaluated for polar method (Kaufmann, 1975).
Polar Index is qualified for measure of ultrafuzziness in 3D polar domain with these conditions.
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 109

Figure 14. Basic rules for construction 3D polar fuzzy type II memebership function.


Figure 15. Three dimensional polar domain of FOU for 3D fuzzy sets, (3D PFOU)

Fuzzy Controllers Recent Advances in Theory and Applications 110





Figure 16. Example of Three polar FOU for 3D fuzzy sets, (3D PFOU) using proposed algorithm
1. IF p
(p,0)
consider as a type I polar fuzzy set Then p
u(p,0)
= p
L(p,0)
AND y(P) = u

( )
( )
1
2
( , ) ( , )
1
0 0
1
( ) ( ) 0 [1, ).
q r
q
U L
q
P g g for q
Area
t
u u

| |
(
|
( = = e
|
(
|

|
\ .
} }

(44)
2. IF
( ) ( ) , ,
1
u L u u
= (high ambiguity) Then
( )
1 P =

( )
( )
1
2
( , ) ( , )
1
0 0
1
( ) ( ) 1 [1, ).
q r
q
U L
q
P g g for q
Area
t
u u

| |
(
|
( = = e
|
(
|

|
\ .
} }

(45)
3.
( )
`
( )
P P =

. Where
`
( )
P is type II fuzzy set and its complement set can be determined
by 1-
( )
, u u
and 1-
( )
, L u
, therefore complement set defined as follow
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 111

`
( ) ( )
( )( )
( )( )
1
3 ,

3 ,

, ,
1
(3 )( , )

(3 )( , )


1

, , | ,

1

h
h
DPRT
U
h
DPRT
u u
h
h
DPRT
L
h
DPRT
v
H
max v
P x
v
H
max v
u
u
u u
u
u


| |
(
|

(

|
+
|
(
|
(

\ .


=
`


| |
(

|

+
|
(
|

\ .

)
(46)
Where 0
U L
H H = = , for the complement set, the ultrafuzziness y is equal:

`
( )
( )
( ) )
1
2
( , ) ( , )
1
0 0
1
1 ( ) 1 ( ) 1,
q r
q
U L
q
P g g P for q
Area
t
u u

| |
(
|
( = + = e
|

(
|

|
\ .
} }


(47)
4. IF
( , ) ( , )
3 3
d c
DPFOU DPFOU
u
< Then
( ) ( )
( , ) ( , ) d c
P P
u
<


.
4.2.2. Finding the optimum interval 3D PMF
The general approach for 3DMF based on upper and lower MF is equal:

( )
( ) ( ) ( )
, ,
,
, 2 , ,
2
P L P U P
IT PMF g OR g
u u
u
u
(
( (
( = = +

(

(48)
Where
( )
P
is polar ultra fuzzy coefficient and
( )
, u

is ultra fuzzy value for

( )
, u u
and
( )
, L u
, in upper and lower bound. For example, Fig.16 presents the principle polar
memberships function of the interval type 2 polar fuzzy sets in position
( )
, u .
5. Polar fuzzy type-2 approximate reasoning
In this part, basic theory of fuzzy polar rule interpolation in fuzzy rule based will be
presented for polar memberships function of type -2 fuzzy sets. Lets us show polar fuzzy
rules with multiple antecedent and single consequent based on T2 PMF rules, Multi Input
Single Output (MISO):
Rule 1: If
1
X is
11
( ) PA
=
&
2
X is
12
( ) PA
=
& &
m
X is
1
( )
m
PA
=
Then Y is
1
( ) PC
=

Rule 2: If
1
X is
21
( ) PA
=
&
2
X is
22
( ) PA
=
& & X
m
is (PA

)
2m
Then Y is
2
( ) PC
=


Fuzzy Controllers Recent Advances in Theory and Applications 112
Rule n: If
1
X is
1
( )
n
PA
=
&
2
X is
2
( )
n
PA
=
& &
m
X is ( )
nm
PA
=
Then Y is

( )
n
PC
=

Where X
p
denotes the pth T2-PFM antecedent and Y denote the T2- PFM consequence.
( )
nm
PA
=
is the n,m th consequence of T2-PFM fuzzy set of Rule n. According polar
MISO method, fuzzy interpolative polar reasoning result which denoted by

P
=
| |
|
|
\ .
can be extracted based on observation polar fuzzy set O
=
| |
|
|
\ .
.
Observation: If
1
X is
1
( ) O
=
& X
2
is
2
( ) O
=
& &
m
X is ( )
m
O
=

Conclusions: O
=
is P
=
.
This method at the first glance, is similar to method which proposed by Chen & Chang
(2011) under the title of fuzzy rule interpolation based on principle membership functions
and uncertainty grad function of interval type-2 fuzzy sets; however, the main difference
between our proposed method with their method relay in type decreasing (Chen & Chang,
2011). Based on Chen & Chang (2011) method first type -2 reduced by type-1, then MISO
applied, but in our method first MISO applied separately on upper and lower. After that the
type-2 polar reasoning result which denoted by P
=
| |
|
|
\ .
extracted and interval 3D PMF
calculated based on section 4.2 theories. This method can be expanded for Multi Input Multi
Output (MIMO) systems based on polar T2 PMFs. For example lets us show polar fuzzy
rules with multiple antecedents and multiple consequent based on T2 PMF rules, Multi
Input Multi Output (MIMO):
Rule 1: If X
1
is
11
( ) PA
=
& &
m
X is
1
( )
m
PA
=
Then Y is
11
( ) PC
=
&
12
( ) PC
=
& &
1
( )
r
PC
=

Rule 2: If
1
X is
21
( ) PA
=
& &
m
X is
2
( )
m
PA
=
Then Y is
21
( ) PC
=
&
22
( ) PC
=
& &
2
( )
r
PC
=

Rule n: If X
1
is
1
( )
n
PA
=
& & X
m
is ( )
nm
PA
=
Then Y is
1
( )
n
PC
=
&
2
( )
n
PC
=
& & ( )
nr
PC
=

Observation: If
1
X is
1
( ) O
=
& &
m
X is ( )
r
O
=

Conclusions: O
=
is
1
P
=
&
2
P
=
& &
r
P
=
.
where, such as polar MISO, X
p
denotes the pth T2-PFM antecedent and Y denote the
T2- PFM consequence which expanded in( )
nr
PC
=
, r denote the number of consequence in
rule n.
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 113
( )
nm
PA
=
is the n,m th consequence of T2-PFM fuzzy set of Rule n. According polar MIMO
method, fuzzy interpolative polar reasoning result which denoted by P
r
| | =
|
|
\ .
can be extracted
based on observation polar fuzzy set O
=
| |
|
|
\ .
in window r for conscience r.
Logical and Mamdani, in the linguistic models considered as prepositions. The general form
of these linguistic rules shows as:
Mamdani:
If X
1
is
1 n
PA
=
| |
|
|
\ .
& &
m
X is
nm
PA
=
| |
|
|
\ .
Then Y is
1 n
PC
=
| |
|
|
\ .
&
2 n
PC
=
| |
|
|
\ .
& &
nr
PC
=
| |
|
|
\ .

Logical:
If X
1
is

1 n
PA
| |
|
|
|
\ .
1
i +
i +
i +
i +
i +
i +
i +
i +
i +
j j
& & X
m
is

nm
PA
| |
|
|
|
\ .
1
i +
i +
i +
i +
i +
i +
i +
i +
i +
j j
Then Y is
1 n
PC
=
| |
|
|
\ .
&
2 n
PC
=
| |
|
|
\ .
& &
nr
PC
=
| |
|
|
\ .

Where
nm
PA
=
| |
|
|
\ .
is the T2 PFM as an antecedent variable in polar system
nr
PC
=
| |
|
|
\ .
is the
consequent variable, n denotes to number of rule and r denoted to r denote the number of
consequences and

1 n
PA
| |
|
|
|
\ .
1
i +
i +
i +
i +
i +
i +
i +
i +
i +
j j
is complement of the T2 PFM as an antecedent variable in
polar system. Concrete effective model proposed by Yager, is a complex method for
combination of these two models, presented as follow:
( ) (1 )( )
y L M
P P P | |
= = =
= + (49)
Where
y
P
=
is the Yager result(complex model), ( )
L
P
=
is result of polar type-2 fuzzy under
logical model,
M
P
=
is consequents of mamdani model and [ is control factor which move
from 0 to 1. (Fazel et al. 2009). The defuzzication method for every 0 must be computed
based on theory which presented in section 4.2. Defuzzification agent in polar system
constructed as three steps,
Step 1. Calculate the center of area in0, in which 0 defined from 0, to 2n (Gold Veins Root)
Step 2. Plot the fuzzy Gold veins determined from first step.
Step 3. Calculate the center of Gold veins and consider it as defuzzify conclusion result.

Fuzzy Controllers Recent Advances in Theory and Applications 114
In step 1, center of area or center of gravity can be calculated in polar system using eq.38,
which is a most common model used :

( )
( )
( )
( )
( )
( )


, ,
0
1
. . .
r
U L
C g g d
Area
u u
u
u
| | (
| ( =
|
(
\ .
}
(50)
and

( )
( )
( )
( )
( )
( )


, ,
0
1
. .
r q q
U L
q m
C g g d
Area
u u
u
u
( | |
| | | | ( |
=
| |
( |
\ . \ .
|
(
\ .
}
(51)
Gold Veins Root is a vector consist of paired
( ) ( )
, C u u , which is shows the direction center
of gravity and it present a useful information from membership function variance and crisp
result without type reduction.
( )
m
C u is a modified Gold Veins Root that can be control the
final defuzzication result, in which q play a defuzzifer role. An example of Gold Veins Root
extracted from polar T2 PMF is depicted in Fig.17. Two result from original and modified
center of gravity by q=3, present in Fig 17.

Figure 17. Gold Veins Root in T2 PMF
In order to effectively implementation and test proposed model in some applications such as
prediction problem and pattern recognition, such as the other models in fuzzy, we need set
in motion to product a crisp result. We chose a heuristic function to generate a best crisp
defuuzy value from Gold Veins Root, based on Eq.51

( )
max min( )
2
GVR GVR
C
(

( =
(

(52)
For the case Fig.18, C=96.33 worth in blurred section from 200, 400 = , provided good
prediction orientation and radios for extension of T2 PFM. Now we present a logical method
for type -2 in 3D techniques. Remove type reduction; turn T2 PMF as faster than existing
techniques and to be more accurate model for type-2 inference techniques because of
combination Logical & Mamdani models.
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 115

Figure 18. a) Input T2 PMF longitudinal cracking, b)Final difuzzification points longitudinal Cracking,
c) Input T2 PMF transverse cracking, d)Final difuzzification points transverse Cracking, e) Input T2PMF
diagonal cracking, f)Final difuzzification points diagonal Cracking, g) Input T2 PMF block cracking,
h)Final difuzzification points block Cracking, i) Input T2 PMF aligator cracking, j)Final difuzzification
points alligator Cracking.

Fuzzy Controllers Recent Advances in Theory and Applications 116
6. Conclusions
In this chapter the basic concepts of new fuzzy sets, three dimensional (3D) memberships
and how they are applied in the design of type-1 and type-2 fuzzy thresholding in control
systems are presented. The robustness of a system highly depends on automatic
fuzzification and membership functions shape and defuzzification. The related
methodology and theoretical base are discussed, using real examples in automatic control in
civil engineering. Selection of a supper membership function is a golden key in fuzzy
controls. A robust method to consider the uncertainty of membership values by using
flexible thresholding for controller problems proposed in the special a polar domain
presented in this chapter. Different fuzzy membership functions may have various impacts
on the systems and, then, different thresholds in control problems. To solve this problem,
type II fuzzy thresholding is recommended. The upper and lower membership functions
promote this dilemma; however the figure of uncertainty (FOU) has a fixed value that is
equal to one, in all the upper and lower membership function. Type-2 fuzzy logic can
effectively improve the control characteristic by using FOU of the membership functions.
A new fuzzy thresholding (flexible thresholding) technique developed, which processes
threshold as a flexible type-2 fuzzy sets. Experimental results are provided in order to
demonstrate the usefulness of the proposed approach. A review of types of fuzzy threshold
methods in control problems provided and their algorithms presented. In type-2 thresholding
method, measurement of fuzziness gives a quantitative index to vagueness. To quantify the
object fuzziness, a suitable membership function based on thresholding for control problems
introduced. A measure for ultra-fuzziness in 3D fuzzy model is proposed. A new method for
thresholding algorithm based on 3D type-2 fuzzy and selection the optimum thresholding in
3D surface are addressed. By an example the validity of novel fuzzy algorithm in control
systems, based on three dimensional membership functions demonstrated.
This paper presents a new type of fuzzy membership functions and uncertainly grade in the
frame of polar systems. The proposed method can be used and generalized for several
problems; however in this paper we present implementation of polar fuzzy type-2(PFT2) as
a part of Hybrid expert system for pavement distress detection and classification.
Vast applications are predicted this fuzzy reasoning. The central idea of this work was to
introduce the application of polar type II fuzzy sets.
The most important aspect of the proposed model is the ability of self-organization of the
membership function and initial height platform without requiring programming.
Additional experiments reinforced this conclusion. More extensive investigations on other
measures of ultrafuzziness and the effect of parameters influencing the width/length of FOU
should certainly be conducted.
Author details
M.H. Fazel Zarandi
Department of Industrial Engineering, Amirkabir University of Technology, Tehran, Iran
A Type-2 Fuzzy Model Based on Three Dimensional
Membership Functions for Smart Thresholding in Control Systems 117
Fereidoon Moghadas Nejad and Hamzeh Zakeri
Department of Civil and Environmental Engineering, Amirkabir University of Technology, Tehran,
Iran
7. References
A Mora, P.M.Vieira, Manivannan.A and Fonseca.J.M, Automated Drusen Detection in
Retinal Images using Analytical Modelling Algorithms, BioMedical Engineering OnLine
2011,
A.Kaufmann, Introduction to the Theory of Fuzzy SubsetsFundamental Theoretical Elements
vol. 1, Academic Press, New York (1975).
C. Hwang, F. Rhee, An interval type-2 fuzzy C spherical shells algorithm, in: Proceedings of
the 2004 IEEE International Conference on Fuzzy Systems, 2004, pp. 11171122.
C. Hwang, F. Rhee, Uncertain fuzzy clustering: interval type-2 fuzzy approach to C-means,
IEEE Transactions on Fuzzy Systems 15 (1) (2007) 107120.
C. Yang, N. Bose, Generating fuzzy membership function with self-organizing feature map,
Pattern Recognition Letters 27 (5) (2006) 356365.
Choi, B.I., Rhee,C.H. Interval type-2 fuzzy membership function generation methods for
pattern recognition, Information Sciences 179 (2009) 21022122.
D.Culpin: Calculation of cubic smoothing splines for equally spaced data. Numerische
Mathematik 1986, 48:627-638.
F. Rhee, B. Choi, Interval type-2 fuzzy membership function design and its application to
radial basis function neural networks, in: Proceedings of the 2007 IEEE International
Conference on Fuzzy Systems, 2007, pp. 20472052.
F. Rhee, C. Hwang, A type-2 fuzzy C-means clustering algorithm, in: Proceedings of the
2001 Joint Conference IFSA/NAFIPS, 2001, pp. 19191926.
F. Rhee, C. Hwang, An interval type-2 fuzzy K-nearest neighbor, in: Proceedings of the 2003
IEEE International Conference on Fuzzy Systems, 2003, pp.802807.
F. Rhee, C. Hwang, An interval type-2 fuzzy perceptron, in: Proceedings of the 2002 IEEE
International Conference on Fuzzy Systems, 2002, pp. 13311335.
F. Rhee, Uncertain fuzzy clustering: insights and recommendations, IEEE Computational
Intelligence Magazine 2 (1) (2007) 4456.
F.Moghadas Nejad, , H.Zakeri, , 2011b. An optimum feature extraction method based on
wavelet radon transform and dynamic neural network for pavement distress
classification. Expert Syst. Appl. 38, 9442 9460.
F.Moghadas Nejad, H.Zakeri, 2011c. A comparison of multiresolution methods for detection
and isolation of pavement distress. Expert Syst.Appl.38,2857-2872.
F.Moghadas Nejad, H.Zakeri, 2011a. An expert system based on wavelet transform and
radon neural network for pavement distress classification .Expert Syst. Appl. 38,7088
7101.
H. Hagras, A hierarchical type-2 fuzzy logic control architecture for autonomous mobile
robots, IEEE Transactions on Fuzzy Systems 12 (4) (2004) 524539.

Fuzzy Controllers Recent Advances in Theory and Applications 118
H.R Tizhoosh., previous Image thresholdingnext using type-2 fuzzy sets, Pattern
Recognition 38 (2005), pp. 23632372.
H.R.Tizhoosh, G. Krell, T. Lilienblum, C.J. Moore and B. Michaelis, Enhancement and
associative restoration of electronic portal images in radiotherapy, Int. J. Med.
Informatics 49 (1998) (2), pp. 157171.
Ioannis K. Vlachos, George D. Sergiadis, Comment on: Image thresholding using type II
fuzzy sets, Pattern Recognition, Volume 41, Issue 5, May 2008, Pages 1810-1811.
J.M.Mendel, Uncertain Rule-Based Fuzzy Logic Systems, Prentice-Hall, Englewood Cliffs,
NJ (2001).
M. Makrehchi, O. Basir, M. Karnel, Generation of fuzzy membership function using
information theory measures and genetic algorithm, Lecture Notes in Computer Science
275 (2003) 603610.
M.H.Fazel Zarandi, M.Zarinbal, I.B.Trksen: Type-II Fuzzy Possibilistic C-Mean Clustering.
IFSA/EUSFLAT Conf. 2009: 30-35.
M.Qiguang, L.Juan, L.Weisheng, S.Junjie, W.Yiding, Three novel invariant moments based
on radon and polar harmonic transforms, Optics Communications 285 (2012) 10441048.
N. Karnik, J. Mendel, Applications of type-2 fuzzy logic systems to forecasting of time series,
Information Sciences 120 (1999) 89111
P.Ensafi, , H.R.Tizhoosh, Type-2 Fuzzy image enhancement International Conference on
Image Analysis and Recognition- ICIAR2005, Springer Lecture Notes in Computer
Science series (Springer LCNS), Toronto, Canada, Sep. 28-30 (2005).
Q. Liang, J. Mendel, Interval type-2 fuzzy logic systems: Theory and Design, IEEE
Transactions on Fuzzy Systems 8 (5) (2000) 535550.
Q. Liang, J. Mendel, Overcoming time-varying co-channel interference using type-2 fuzzy
adaptive lter, IEEE Transactions on Circuits and Systems-II:Analog and Digital Signal
Processing 47 (2000) 14191428.
R. John, Embedded interval valued type-2 fuzzy sets, in: Proceedings of the 2002 IEEE
International Conference on Fuzzy Systems, (2002), pp. 13161320.
R.N Pal and .C Bezdek., measuring fuzzy uncertainty, IEEE , transactions on fuzzy systems,
, 1994,vol. 2, no. 2, PP.107118.
S. Medasani, J. Kim, R. Krishnapuram, An overview of membership function generation
techniques for pattern recognition, International Journalof Approximate Reasoning 19
(34) (1998) 391417.
S. Wang, Generating fuzzy membership functions: a monotonic neural network, Fuzzy Sets
and Systems 61 (1) (1994) 7181.
S.K. Pal and C.A. Murthy, Fuzzy thresholding: mathematical framework bound functions
and weighted moving average technique, Pattern Recognition Lett. 11 (1990), pp. 197
206.
Z.Daqi, S,Qu. L,He. and S.Sh. Automatic Ridgelet image enhancement algorithm for road
crack image based on fuzzy entropy and fuzzy divergence. (2009),In press, Optics and
Lasers in Engineering.
Chapter 0
Fuzzy Control of Nonlinear Systems
with General Performance Criteria
Xin Wang, Edwin E. Yaz, James Long and Tim Miller
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/48298
1. Introduction
Research on control of non-linear systems over the years has produced many results: control
based on linearization, global feedback linearization, non-linear H

control, sliding mode

control, variable structure control, state dependent Riccati equation control, etc [5]. This
chapter will focus on fuzzy control techniques. Fuzzy control systems have recently shown
growing popularity in non-linear system control applications. A fuzzy control system is
essentially an effective way to decompose the task of non-linear system control into a group
of local linear controls based on a set of design-specic model rules. Fuzzy control also
provides a mechanism to blend these local linear control problems all together to achieve
overall control of the original non-linear system. In this regard, fuzzy control technique has its
unique advantage over other kinds of non-linear control techniques. Latest research on fuzzy
control systems design is aimed to improve the optimality and robustness of the controller
performance by combining the advantage of modern control theory with the Takagi-Sugeno
fuzzy model [710, 13, 14].
In this chapter, we address the non-linear state feedback control design of both
continuous-time and discrete-time non-linear fuzzy control systems using the Linear Matrix
Inequality (LMI) approach. We characterize the solution of the non-linear control problem
with the LMI, which provides a sufcient condition for satisfying various performance
criteria. A preliminary investigation into the LMI approach to non-linear fuzzy control
systems can be found in [7, 8, 13]. The purpose behind this novel approach is to convert
a non-linear system control problem into a convex optimization problem which is solved
by a LMI at each time. The recent development in convex optimization provides efcient
algorithms for solving LMIs. If a solution can be expressed in a LMI form, then there exist
optimization algorithms providing efcient global numerical solutions [3]. Therefore if the
LMI is feasible, then LMI control technique provides globally stable solutions satisfying the
corresponding mixed performance criteria [4, 6, 1520]. We further propose to employ mixed
performance criteria to design the controller guaranteeing the quadratic sub-optimality with
inherent stability property in combination with dissipative type of disturbance attenuation.
Chapter 5
2 Fuzzy Controllers
In the following sections, we rst introduce the Takagi-Sugeno fuzzy modelling for non-linear
systems in both continuous time and discrete time. We then propose the general performance
criteria in section 3. Then, the LMI control solutions are derived to characterize the
optimal and robust fuzzy control of continuous time and discrete time non-linear systems,
respectively. The inverted pendulum system control is used as an illustrative example to
demonstrate the effectiveness and robustness of our proposed approaches.
The following notation is used in this work: x R
n
denotes n-dimensional real vector with
norm x = (x
T
x)
1/2
where (.)
T
indicates transpose. A 0 for a symmetric matrix denotes
a positive semi-denite matrix. L
2
and l
2
denotes the space of innite sequences of nite
dimensional random vectors with nite energy, i.e.

0
x
t

2
< in continuous-time, and

k=0
x
k

2
< in discrete-time, respectively.
2. Takagi-Sugeno system model
The importance of the Takagi-Sugeno fuzzy system model is that it provides an effective way
to decompose a complicated non-linear system into local dynamical relations and express
those local dynamics of each fuzzy implication rule by a linear system model. The overall
fuzzy non-linear system model is achieved by fuzzy blending of the linear system models,
so that the overall non-linear control performance is achieved. Both of the continuous-time
and the discrete-time system models are summarized below.
2.1. Continuous-time Takagi-Sugeno system model
The i
th
rule of the Takagi-Sugeno fuzzy model can be expressed by the following forms:
Model Rule i:
If
1
(t) is M
i1
,
2
(t) is M
i2
,... and
p
(t) is M
i p
,
Then the input-afne continuous-time fuzzy system equation is:
x(t) = A
i
x(t) + B
i
u(t) + F
i
w(t)
y(t) = C
i
x(t) + D
i
u(t) + Z
i
w(t)
i = 1, 2, 3, ..., r (1)
where x(t) R
n
is the state vector, u(t) R
m
is the control input vector, y(t) R
q
is
the performance output vector, w(t) R
s
is L
2
type of disturbance, r is the total number of
model rules, M
ij
is the fuzzy set. The coefcient matrices are A
i
R
nn
, B
i
R
nm
, F
i

R
ns
, C
i
R
qn
, D
i
R
qm
, Z
i
R
qs
. And
1
, ...,
p
are known premise variables, which
can be functions of state variables, external disturbance and time.
It is assumed that the premises are not the function of the input vector u(t), which is needed
to avoid the defuzzication process of fuzzy controller. If we use (t) to denote the vector
containing all the individual elements
1
(t),
2
(t), ...,
p
(t), then the overall fuzzy system is
x(t) =

r
i=1
g
i
((t))[A
i
x(t) + B
i
u(t) + F
i
w(t)]

r
i=1
g
i
((t))
=
r

i=1
h
i
((t))[A
i
x(t) + B
i
u(t) + F
i
w(t)]
y(t) =

r
i=1
g
i
((t))[C
i
x(t) + D
i
u(t) + Z
i
w(t)]

r
i=1
g
i
((t))
=
r

i=1
h
i
((t))[C
i
x(t) + D
i
u(t) + Z
i
w(t)] (2)
120 Fuzzy Controllers Recent Advances in Theory and Applications
Fuzzy Control of Nonlinear Systems with General Performance Criteria 3
where
(t) = [
1
(t),
2
(t), ...,
p
(t)] (3)
g
i
((t)) =
p

j=1
M
ij
(
j
(t)) (4)
h
i
((t)) =
g
i
((t))

r
i=1
g
i
((t))
(5)
for all time t. The term M
ij
(
j
(t)) is the grade membership function of
j
(t) in M
ij
.
Since, the following properties hold
r

i=1
g
i
((t)) > 0
g
i
((t)) 0, i = 1, 2, 3, ..., r (6)
We have
r

i=1
h
i
((t)) = 1
h
i
((t)) 0, i = 1, 2, 3, ..., r (7)
for all time t.
It is assumed that the state feedback is available and the non-linear state feedback control
input is given by
u(t) =
r

i=1
h
i
((t))K
i
x(t) (8)
Substituting this into the system and performance output equation, we have
x(t) =
r

i=1
r

j=1
h
i
((t))h
j
((t))(A
i
B
i
K
j
)x(t) +
r

i=1
h
i
((t))F
i
w(t)
y(t) =
r

i=1
r

j=1
h
i
((t))h
j
((t))(C
i
D
i
K
j
)x(k) +
r

i=1
h
i
((t))Z
i
w(t) (9)
Using the notation
G
ij
= A
i
B
i
K
j
H
ij
= C
i
D
i
K
j
(10)
then the system equation becomes
x(t) =
r

i=1
r

j=1
h
i
((t))h
j
((t))G
ij
x(t) +
r

i=1
h
i
((t))F
i
w(t)
y(t) =
r

i=1
r

j=1
h
i
((t))h
j
((t))H
ij
x(t) +
r

i=1
h
i
((t))Z
i
w(t) (11)
121 Fuzzy Control of Nonlinear Systems with General Performance Criteria
4 Fuzzy Controllers
2.2. Discrete-time Takagi-Sugeno system model
At time step k, the i
th
rule of the Takagi-Sugeno fuzzy model can be expressed by the following
forms:
Model Rule i:
If
1
(k) is M
i1
,
2
(k) is M
i2
,... and
p
(k) is M
i p
,
Then the input-afne discrete-time fuzzy system equation is:
x(k +1) = A
i
x(k) + B
i
u(k) + F
i
w(k)
y(k) = C
i
x(k) + D
i
u(k) + Z
i
w(k)
i = 1, 2, 3, ..., r (12)
where x(k) R
n
is the state vector, u(k) R
m
is the control input vector, y(k) R
q
is
the performance output vector, w(k) R
s
is l
2
type of disturbance, r is the total number of
model rules, M
ij
is the fuzzy set. The coefcient matrices are A
i
R
nn
, B
i
R
nm
, F
i

R
ns
, C
i
R
qn
, D
i
R
qm
, Z
i
R
qs
. And
1
, ...,
p
are known premise variables which
can be functions of state variables, external disturbance and time.
It is assumed that the premises are not the function of the input vector u(k), which is needed
to avoid the defuzzication process of fuzzy controller. If we use (k) to denote the vector
containing all the individual elements
1
(k),
2
(k), ...,
p
(k), then the overall fuzzy system is
x(k+1) =

r
i=1
g
i
((k))A
i
x(k)+B
i
u(k)+F
i
w(k)

r
i=1
g
i
((k))
=
r

i=1
h
i
((k))A
i
x(k)+B
i
u(k)+F
i
w(k)
y(k) =

r
i=1
g
i
((k))C
i
x(k)+D
i
u(k)+Z
i
w(k)

r
i=1
g
i
((k))
=
r

i=1
h
i
((k))C
i
x(k)+D
i
u(k)+Z
i
w(k) (13)
where
(k) = [
1
(k),
2
(k), ...,
p
(k)] (14)
g
i
((k)) =
p

j=1
M
ij
(
j
(k)) (15)
h
i
((k)) =
g
i
((k))

r
i=1
g
i
((k))
(16)
for all k. The term M
ij
(
j
(k)) is the grade membership function of
j
(k) in M
ij
.
Since, the following properties hold
r

i=1
g
i
((k)) > 0
g
i
((k)) 0, i = 1, 2, 3, ..., r (17)
We have
r

i=1
h
i
((k)) = 1
h
i
((k)) 0, i = 1, 2, 3, ..., r (18)
122 Fuzzy Controllers Recent Advances in Theory and Applications
Fuzzy Control of Nonlinear Systems with General Performance Criteria 5
for all k.
It is assumed that the state feedback is available and the non-linear state feedback control
input is given by
u(k) =
r

i=1
h
i
((k))K
i
x(k) (19)
Substituting this into the system and performance output equation, we have
x(k +1) =
r

i=1
r

j=1
h
i
((k))h
j
((k))(A
i
B
i
K
j
)x(k) +
r

i=1
h
i
((k))F
i
w(k)
y(k) =
r

i=1
r

j=1
h
i
((k))h
j
((k))(C
i
D
i
K
j
)x(k) +
r

i=1
h
i
((k))Z
i
w(k) (20)
Using the notation
G
ij
= A
i
B
i
K
j
H
ij
= C
i
D
i
K
j
(21)
then the system equation becomes
x(k +1) =
r

i=1
r

j=1
h
i
((k))h
j
((k))G
ij
x(k) +
r

i=1
h
i
((k))F
i
w(k)
y(k) =
r

i=1
r

j=1
h
i
((k))h
j
((k))H
ij
x(k) +
r

i=1
h
i
((k))Z
i
w(k) (22)
3. General performance criteria
In this section, we propose the general performance criteria for non-linear control design,
which yields a mixed Non-Linear Quadratic Regular (NLQR) in combination with H

or
dissipative performance index. The commonly used system performance criteria, including
bounded-realness, positive-realness, sector boundedness and quadratic cost criterion, become
special cases of the general performance criteria. Both the continuous-time and discrete-time
general performance criteria are given below:
3.1. Continuous-time general performance criteria
Consider the quadratic Lyapunov function
V(t) = x
T
(t)Px(t) > 0 (23)
for the following difference inequality

V(t) + x
T
(t)Qx(t) +u
T
(t)Ru(t) +y
T
(t)y(t) y
T
(t)w(t) +w
T
(t)w(t) 0 (24)
with Q > 0, R > 0 functions of x(t).
123 Fuzzy Control of Nonlinear Systems with General Performance Criteria
6 Fuzzy Controllers
Note that upon integration over time from 0 to T
f
, (24) yields
V(T
f
) +

T
f
0
[(x
T
(t)Qx(t) +u
T
(t)Ru(t)]dt +

T
f
0
[y
T
(t)y(t) y
T
(t)w(t) +w
T
(t)w(t)]dt V(0) (25)
By properly specifying the value of weighing matrices Q, R, C
i
, D
i
, Z
i
and , , , mixed
performance criteria can be used in non-linear control design, which yields a mixed
Non-linear Quadratic Regulator (NLQR) in combination with dissipative type performance
index with disturbance reduction capability. For example, if we take = 1, = 0, < 0, (25)
yields
V(T
f
) +

T
f
0
[(x
T
(t)Qx(t) +u
T
(t)Ru(t) +y
T
(t)y(t)]dt +
V(0)

T
f
0
[w
T
(t)w(t)]dt (26)
which is a mixed NLQR H

Design [1618].
Other possible performance criteria which can be used in this framework with various design
parameters , , are given in Table.1. Design coefcients and can be maximized or
minimized to optimize the controller behavior. It should also be noted that the satisfaction of
any of the criteria in Table 1 will also guarantee asymptotic stability of the controlled system.
3.2. Discrete-time general performance criteria
Consider the quadratic Lyapunov function
V(k) = x
T
(k)Px(k) (27)
for the following difference inequality
V(k +1) V(k) + x
T
(k)Qx(k) +u
T
(k)Ru(k) +y
T
(k)y(k) y
T
(k)w(k) +w
T
(k)w(k) 0
(28)
with Q > 0, R > 0 functions of x(k).
Note that upon summation over k, (28) yields
V(N) +
N1

k=0
(x
T
(k)Qx(k) +u
T
(k)Ru(k) +y
T
(k)y(k) y
T
(k)w(k) +w
T
(k)w(k)) V(0)
(29)
By properly specifying the value of weighing matrices Q, R, C
i
, D
i
, Z
i
and , , , mixed
performance criteria can be used in non-linear control design, which yields a mixed
Non-linear Quadratic Regulator (NLQR) in combination with dissipative type performance
index with disturbance reduction capability. For example, if we take = 1, = 0, < 0, (29)
yields
V(N) +
N1

k=0
(x
T
(k)Qx(k) +u
T
(k)Ru(k) +y
T
(k)y(k)) V(0)
N1

k=0
w
T
(k)w(k) (30)
124 Fuzzy Controllers Recent Advances in Theory and Applications
Fuzzy Control of Nonlinear Systems with General Performance Criteria 7
which is a mixed NLQRH

Design [1618]. In (19), can be minimized to achieve a smaller

l
2
l
2
or H

gain for the closed loop system.

Other possible performance criteria which can be used in this framework with various design
parameters , , are given in Table.1. Design coefcients and can be maximized or
minimized to optimize the controller behavior. It should also be noted that the satisfaction of
any of the criteria in Table 1 will also guarantee asymptotic stability of the controlled system.
Performance Criteria
1 0 <0 NLQR H

Design
0 1 0 NLQR Passivity Design
0 1 >0 NLQR Input Strict Passivity Design
>0 1 0 NLQR Output Strict Passivity Design
>0 1 >0 NLQR Very Strict Passivity
Table 1. Various performance criteria in a general framework
4. Fuzzy LMI control of continuous time non-linear systems with general
performance criteria
The main results of this chapter are summarized in section 4 and section 5. The following
theorem provides the fuzzy LMI control to the continuous time non-linear systems with
general performance criteria.
Theorem 1 Given the system model and performance output (2) and control input (8), if there
exist matrices S = P
1
> 0 for all t 0, such that the following LMI holds:

11

12

13

14

15

22

23
0 0
I 0 0
R
1
0
I

0 (31)
where

11
=
1
2
[SA
T
i
M
j
B
T
i
+SA
T
j
M
T
i
B
T
j
+ A
i
S B
i
M
j
+ A
j
S B
j
M
i
]

12
=
1
2
(F
i
+ F
j
) +

4
[SC
T
i
M
j
D
T
i
+SC
T
j
M
T
i
D
T
j
]

13
=
1
2

1/2
[SC
T
i
M
j
D
T
i
+SC
T
j
M
T
i
D
T
j
]

14
=
1
2
(M
T
i
+ M
T
j
)

15
= SQ
T/2

22
= I +
1
2
(Z
i
+ Z
j
)
T

23
=
1
2

1/2
[Z
i
+ Z
j
]
T
(32)
125 Fuzzy Control of Nonlinear Systems with General Performance Criteria
8 Fuzzy Controllers
using the notation
M
i
= K
i
P
1
= K
i
S (33)
then inequality (24) is satised.
Proof
By applying system model and performance output (2)(11), and state feedback input (8), the
performance index inequality (24) becomes
[
r

i=1
r

j=1
h
i
((t))h
j
((t))G
ij
x(t) +
r

i=1
h
i
((t))F
i
w(t)]
T
Px(t) +
x
T
(t)P[
r

i=1
r

j=1
h
i
((t))h
j
((t))G
ij
x(t) +
r

i=1
h
i
((t))F
i
w(t)] +
x
T
(t)Qx(t) + [
r

i=1
h
i
(t)K
i
x(t)]
T
R[
r

i=1
h
i
(t)K
i
x(t)]
[
r

i=1
r

j=1
h
i
((t))h
j
((t))H
ij
x(t) +
r

i=1
h
i
((t))Z
i
w(t)]
T
[
r

i=1
r

j=1
h
i
((t))h
j
((t))H
ij
x(t) +
r

i=1
h
i
((t))Z
i
w(t)]
[
r

i=1
r

j=1
h
i
((t))h
j
((t))H
ij
x(t) +
r

i=1
h
i
((t))Z
i
w(t)]
T
w(t)
+w
T
(t)w(t) 0 (34)
Inequality (34) is equivalent to

x
T
(t) w
T
(t)

11

12

22

x(t)
w(t)

0 (35)
where

11
= (

j
h
i
h
j
G
ij
)
T
P + P(

j
h
i
h
j
G
ij
) + Q+ [

i
h
i
K
i
]
T
R[

i
h
i
K
i
] +
[

j
h
i
h
j
H
ij
]
T
[

j
h
i
h
j
H
ij
]

12
= P(

i
h
i
F
i
) +[

j
h
i
h
j
H
ij
]
T
[

i
h
i
Z
i
]

2
[

j
h
i
h
j
H
ij
]
T

22
= I +[

i
h
i
Z
i
]
T
[

i
h
i
Z
i
] [

i
h
i
Z
i
]
T
(36)
Inequality (35) can be rewritten as

11

12

22

[
i

j
h
i
h
j
H
ij
]
T
[
i
h
i
Z
i
]
T

[
i

j
h
i
h
j
H
ij
] [
i
h
i
Z
i
]

0 (37)
126 Fuzzy Controllers Recent Advances in Theory and Applications
Fuzzy Control of Nonlinear Systems with General Performance Criteria 9
where

11
= (

j
h
i
h
j
G
ij
)
T
P P(

j
h
i
h
j
G
ij
) Q[

i
h
i
K
i
]
T
R[

i
h
i
K
i
]

12
= P(

i
h
i
F
i
) +

2
[

j
h
i
h
j
H
ij
]
T

22
= I + [

i
h
i
Z
i
]
T
(38)
By applying Schur complement to inequality (37), we have

11

12

1/2
[
i

j
h
i
h
j
H
ij
]
T

22

1/2
[
i
h
i
Z
i
]
T
I

0 (39)
Similarly, inequality (39) can also be written as

11

12

1/2
[
i

j
h
i
h
j
H
ij
]
T

22

1/2
[
i
h
i
Z
i
]
T
I

[
i
h
i
K
i
]
T
0
0

[
i
h
i
K
i
] 0 0

0 (40)
where

11
= (

j
h
i
h
j
G
ij
)
T
P P(

j
h
i
h
j
G
ij
) Q

12
= P(

i
h
i
F
i
) +

2
[

j
h
i
h
j
H
ij
]
T

22
= I + [

i
h
i
Z
i
]
T
(41)
By applying Schur complement again to (40), we have

11

12

1/2
[
i

j
h
i
h
j
H
ij
]
T
[
i
h
i
K
i
]
T

22

1/2
[
i
h
i
Z
i
]
T
0
I 0
R
1

0 (42)
Equivalently, we have

j
h
i
h
j

11

12

13

14

22

23
0
I 0
R
1

0 (43)
127 Fuzzy Control of Nonlinear Systems with General Performance Criteria
10 Fuzzy Controllers
where

11
=
1
2
[(A
i
B
i
K
j
) + (A
j
B
j
K
i
)]
T
P
1
2
P[(A
i
B
i
K
j
) + (A
j
B
j
K
i
)] Q

12
=
1
2
P(F
i
+ F
j
) +

4
[(C
i
D
i
K
j
) + (C
j
D
j
K
i
)]
T

13
=
1
2

1/2
[(C
i
D
i
K
j
) + (C
j
D
j
K
i
)]
T

14
=
1
2
(K
i
+K
j
)
T

22
= I +
1
2
(Z
i
+ Z
j
)
T

23
=
1
2

1/2
(Z
i
+ Z
j
)
T
(44)
Therefore, we have the following LMI

11

12

13

14

22

23
0
I 0
R
1

0 (45)
By multiplying both sides of the LMI above by the block diagonal matrix diag{S, I, I, I}, where
S = P
1
, and using the notation
M
i
= K
i
P
1
= K
i
S (46)
we obtain

X
11
X
12
X
13
X
14
X
22
X
23
0
I 0
R
1

0 (47)
where
X
11
=
1
2
[SA
T
i
M
j
B
T
i
+SA
T
j
M
T
i
B
T
j
+ A
i
S B
i
M
j
+ A
j
S B
j
M
i
] SQS
X
12
=
1
2
(F
i
+ F
j
) +

4
[SC
T
i
M
T
j
D
T
i
+SC
T
j
M
T
i
D
T
j
]
X
13
=
1
2

1/2
[SC
T
i
M
T
j
D
T
i
+SC
T
j
M
T
i
D
T
j
]
X
14
=
1
2
(M
T
i
+ M
T
j
)
X
22
= I +
1
2
(Z
i
+ Z
j
)
T
X
23
=
1
2

1/2
(Z
i
+ Z
j
)
T
(48)
128 Fuzzy Controllers Recent Advances in Theory and Applications
Fuzzy Control of Nonlinear Systems with General Performance Criteria 11
By applying Schur complement again, the nal LMI is derived

11

12

13

14

15

22

23
0 0
I 0 0
R
1
0
I

0 (49)
where

11
=
1
2
[SA
T
i
M
j
B
T
i
+SA
T
j
M
T
i
B
T
j
+ A
i
S B
i
M
j
+ A
j
S B
j
M
i
]

12
=
1
2
(F
i
+ F
j
) +

4
[SC
T
i
M
j
D
T
i
+SC
T
j
M
T
i
D
T
j
]

13
=
1
2

1/2
[SC
T
i
M
j
D
T
i
+SC
T
j
M
T
i
D
T
j
]

14
=
1
2
(M
T
i
+ M
T
j
)

15
= SQ
T/2

22
= I +
1
2
(Z
i
+ Z
j
)
T

23
=
1
2

1/2
[Z
i
+ Z
j
]
T
(50)
Hence, if the LMI (49) holds, inequality (24) is satised. This concludes the proof of the
theorem.
Remark 1: For the chosen performance criterion, the LMI (49) need to be solved at each time
to nd matrices S, M, by using relation (33), we can nd the feedback control gain, therefore,
the feedback control can be found to satisfy the chosen criterion.
5. Fuzzy LMI control of discrete time non-linear systems with general
performance criteria
This section summarizes the main results for fuzzy LMI control of discrete time non-linear
systems with general performance criteria:
Theorem 2: Given the closed loop system and performance output (13), and control input
(19), if there exist matrices S = P
1
> 0 for all k 0, such that the following LMI holds:

11

12

13

14

15

16

22

23

24
0 0
S 0 0 0
I 0 0
R
1
0
I

0 (51)
129 Fuzzy Control of Nonlinear Systems with General Performance Criteria
12 Fuzzy Controllers
where

11
= S

12
=

4
(C
i
S D
i
Y
j
+C
j
S D
j
Y
i
)
T

13
=
1
2
(A
i
S B
i
Y
j
+ A
j
S B
j
Y
i
)
T

14
=
1
2

1/2
(C
i
S D
i
Y
j
+C
j
S D
j
Y
i
)
T

15
=
1
2
(Y
i
+Y
j
)
T

16
= SQ
T/2

22
= I +

2
(Z
i
+ Z
j
)
T

23
=
1
2

1/2
(F
i
+ F
j
)
T

24
=
1
2

1/2
(Z
i
+ Z
j
)
T
(52)
and
S(k +1) > S(k) (53)
where S(k) = P
1
(k), then (28) is satised with the feedback control gain being found by
K(k) = Y(k)P(k) (54)
Proof
The performance index inequality (28) can be explicitly written as
[
r

i=1
r

j=1
h
i
((k))h
j
((k))G
ij
x(k) +
r

i=1
h
i
((k))F
i
w(k)]
T
P [
r

i=1
r

j=1
h
i
((k))h
j
((k))G
ij
x(k) +
r

i=1
h
i
((k))F
i
w(k)]
x
T
(k)Px(k) + x
T
(k)Qx(k) + [
r

i=1
h
i
((k))K
i
x(k)]
T
R[
r

i=1
h
i
((k))K
i
x(k)] +
[
r

i=1
r

j=1
h
i
((k))h
j
((k))H
ij
x(k) +
r

i=1
h
i
((k))Z
i
w(k)]
T
[
r

i=1
r

j=1
h
i
((k))h
j
((k))H
ij
x(k) +
r

i=1
h
i
((k))Z
i
w(k)]
[
r

i=1
r

j=1
h
i
((k))h
j
((k))H
ij
x(k) +
r

i=1
h
i
((k))Z
i
w(k)]
T
w(k)
+w
T
(k)w(k) 0 (55)
130 Fuzzy Controllers Recent Advances in Theory and Applications
Fuzzy Control of Nonlinear Systems with General Performance Criteria 13
Equivalently,

x
T
(k) w
T
(k)

P + Q 0
0 I

x(k)
w(k)

x
T
(k) w
T
(k)

(
i

j
h
i
h
j
G
ij
) (
i
h
i
F
i
)

T
P

(
i

j
h
i
h
j
G
ij
) (
i
h
i
F
i
)

x(k)
w(k)

+
+x
T
(k)[

i
h
i
K
i
]
T
R[

i
h
i
K
i
]x(k) +

x
T
(k) w
T
(k)

(
i

j
h
i
h
j
H
ij
) (
i
h
i
Z
i
)

(
i

j
h
i
h
j
H
ij
) (
i
h
i
Z
i
)

x(k)
w(k)

x
T
(k) w
T
(k)

(
i

j
h
i
h
j
H
ij
) (
i
h
i
Z
i
)

T
w(k) 0
(56)
which can be written, after collecting terms, as

x
T
(k) w
T
(k)

11

12

22

x(k)
w(k)

x
T
(k) w
T
(k)

(
i

j
h
i
h
j
G
ij
) (
i
h
i
F
i
)

T
P

(
i

j
h
i
h
j
G
ij
) (
i
h
i
F
i
)

x(k)
w(k)

x
T
(k) w
T
(k)

(
i

j
h
i
h
j
H
ij
) (
i
h
i
Z
i
)

(
i

j
h
i
h
j
H
ij
) (
i
h
i
Z
i
)

x(k)
w(k)

0
(57)
where

11
= P Q[

i
h
i
K
i
]
T
R[

i
h
i
K
i
]

12
=

2
[

j
h
i
h
j
H
ij
]
T

22
= I + [

i
h
i
Z
i
]
T
(58)
Equivalently, we have

11

12

22

(
i

j
h
i
h
j
G
ij
) (
i
h
i
F
i
)

T
P

(
i

j
h
i
h
j
G
ij
) (
i
h
i
F
i
)

(
i

j
h
i
h
j
H
ij
) (
i
h
i
Z
i
)

(
i

j
h
i
h
j
H
ij
) (
i
h
i
Z
i
)

0
(59)
By applying Schur complement, we obtain

11

12
(
i

j
h
i
h
j
G
ij
))
T

22
(
i
h
i
F
i
)
T
P
1

(
i

j
h
i
h
j
H
ij
) (
i
h
i
Z
i
)

(
i

j
h
i
h
j
H
ij
) (
i
h
i
Z
i
)

0
(60)
131 Fuzzy Control of Nonlinear Systems with General Performance Criteria
14 Fuzzy Controllers
By applying Schur complement again, we obtain

11

12
(
i

j
h
i
h
j
G
ij
))
T

1/2
(
i

j
h
i
h
j
H
ij
)
T

22
(
i
h
i
F
i
)
T

1/2
(
i
h
i
Z
i
)
T
P
1
0
I

0
(61)
Equivalently, the following inequality holds

11

12
(
i

j
h
i
h
j
G
ij
))
T

1/2
(
i

j
h
i
h
j
H
ij
)
T

22
(
i
h
i
F
i
)
T

1/2
(
i
h
i
Z
i
)
T
P
1
0
I

(
i
h
i
K
i
)
T
0
0
0

(
i
h
i
K
i
) 0 0 0

0
(62)
where

11
= P Q

12
=

2
[

j
h
i
h
j
H
ij
]
T

22
= I + [

i
h
i
Z
i
]
T
(63)
By applying Schur complement one more time, we have

11

12
(
i

j
h
i
h
j
G
ij
))
T

1/2
(
i

j
h
i
h
j
H
ij
)
T
(
i
h
i
K
i
)
T

22
(
i
h
i
F
i
)
T

1/2
(
i
h
i
Z
i
)
T
0
P
1
0 0
I 0
R
1

0
(64)
By factoring out the
i

j
h
i
(
k
)h
j
(
k
) term, we have

11

12

13

14

15

22

23

24
0
P
1
0 0
I 0
R
1

0
(65)
132 Fuzzy Controllers Recent Advances in Theory and Applications
Fuzzy Control of Nonlinear Systems with General Performance Criteria 15
where

11
= P Q

12
=

4
[H
ji
+ H
ij
]
T

13
=
1
2
(G
ji
+ G
ij
))
T

14
=
1
2

1/2
(H
ij
+ H
ji
)
T

15
=
1
2
(K
i
+K
j
)
T

22
= I +

2
(Z
i
+ Z
j
)
T

23
=
1
2
(F
i
+ F
j
)
T

24
=
1
2

1/2
(Z
i
+ Z
j
)
T
(66)
By pre-multiplying and post-multiplying the matrix with the block diagonal matrix
diag(S, I, I, I, I) , where S = P
1
, and applying Schur complement again, the following LMI
result is obtained

11

12

13

14

15

16

22

23

24
0 0
S 0 0 0
I 0 0
R
1
0
I

0 (67)
where

11
= S

12
=

4
(C
i
S D
i
Y
j
+C
j
S D
j
Y
i
)
T

13
=
1
2
(A
i
S B
i
Y
j
+ A
j
S B
j
Y
i
)
T

14
=
1
2

1/2
(C
i
S D
i
Y
j
+C
j
S D
j
Y
i
)
T

15
=
1
2
(Y
i
+Y
j
)
T

16
= SQ
T/2

22
= I +

2
(Z
i
+ Z
j
)
T

23
=
1
2

1/2
(F
i
+ F
j
)
T

24
=
1
2

1/2
(Z
i
+ Z
j
)
T
(68)
133 Fuzzy Control of Nonlinear Systems with General Performance Criteria
16 Fuzzy Controllers
where S(k) = P
1
(k), then (28) is satised with the feedback control gain being found by
K(k) = Y(k)P(k) (69)
6. Application to the inverted pendulum system
The inverted pendulum on a cart problem is a benchmark control problem used widely to test
control algorithms. A pendulum beam attached at one end can rotate freely in the vertical
2-dimensional plane. The angle of the beam with respect to the vertical direction is denoted
at angle . The external force u is desired to set angle of the beam (x
1
) and angular velocity

(x
2
) to zero while satisfying the mixed performance criteria. A model of the inverted
pendulum on a cart problem is given by [1, 9]:
x
1
= x
2
+
1
w
x
2
=
gsin(x
1
) amLx
2
2
sin(2x
1
)/2 acos(x
1
)u
4L/3 amLcos
2
(x
1
)
+
2
w (70)
where x
1
is the angle of the pendulum from vertical direction, x
2
is the angular velocity of the
pendulum, g is the gravity constant, m is the mass of the pendulum, M is the mass of the cart,
L is the length of the center of mass (the entire length of the pendulum beam equals 2L), u is
the external force, control input to the system, w is the L
2
type of disturbance, a =
1
m+M
is a
constant, and
1
.
2
is the weighing coefcients of disturbance.
Due to the system non-linearity, we approximate the system using the following two-rule
fuzzy model:
continuous-time fuzzy model
Rule 1: If |x
1
(t)| is close to zero,
Then x(t) = A
1
x(t) + B
1
u(t) + F
1
w(t)
Rule 2: If |x
1
(t)| is close to /2,
Then x(t) = A
2
x(t) + B
2
u(t) + F
2
w(t)
where
A
1
=

0 1
g
4L/3amL
0

B
1
=

a
4L/3amL

F
1
=

A
2
=

0 1
2g
(4L/3amL
2
)
0

B
1
=

a
4L/3amL
2

F
1
=

with = cos(80
o
) (71)
discrete-time fuzzy model
Rule 1: If |x
1
(k)| is close to zero,
Then x(k +1) = A
1
x(k) +B
1
u(k) +F
1
w(k)
Rule 2: If |x
1
(k)| is close to /2,
Then x(k +1) = A
2
x(k) +B
2
u(k) +F
2
w(k)
134 Fuzzy Controllers Recent Advances in Theory and Applications
Fuzzy Control of Nonlinear Systems with General Performance Criteria 17
where
A
1
=

1 T
gT
4L/3amL
1

B
1
=

aT
4L/3amL

F
1
=

1
T

2
T

A
2
=

1 T
2gT
(4L/3amL
2
)
1

B
2
=

aT
4L/3amL
2

F
2
=

1
T

2
T

with = cos(80
o
), Sampling time T = 0.001 (72)
The following values are used in our simulation:
M = 8kg, m = 2kg, L = 0.5m, g = 9.8m/s
2
,
1
= 1,
2
= 0
and the initial condition of x
1
(0) = /6, x
2
(0) = /6. The membership function of Rule 1
and Rule 2 is shown below in Fig.1.
Figure 1. Membership functions of Rule 1 and Rule 2.
Figure 2. Angle trajectory of the inverted pendulum.
The feedback control gain can be found from (31)(51) by solving the LMI at each time. The
following design parameters are chosen to satisfy:
Mixed NLQR H

criteria:
C = [1 1], D = [1], Q = diag[1001], R = 1, = 1, = 0, = 5
Mixed NLQR passivity criteria:
C = [1 1], D = [1], Q = diag[1001], R = 1, = 1, = 5, = 0
135 Fuzzy Control of Nonlinear Systems with General Performance Criteria
18 Fuzzy Controllers
Figure 3. Angular velocity trajectory of the inverted pendulum.
Figure 4. Control input applied to the inverted pendulum.
The mixed criteria control performance results are shown in the Figs.2-4. From these gures,
we nd that the novel fuzzy LMI control has satisfactory performance. The mixed NLQR
H

criteria control has a smaller overshoot and a faster response than the one with passivity
property. The new technique controls the inverted pendulum very well under the effect of
nite energy disturbance. It should also be noted that the LMI fuzzy control with mixed
performance criteria satises global asymptotic stability.
7. Summary
This chapter presents a novel fuzzy control approach for both of continuous time and discrete
time non-linear systems based on the LMI solutions. The Takagi-Sugeno fuzzy model is
applied to decompose the non-linear system. Multiple performance criteria are used to design
the controller and the relative weighting matrices of these criteria can be achieved by choosing
different coefcient matrices. The optimal control can be obtained by solving LMI at each
time. The inverted pendulum is used as an example to demonstrate its effectiveness. The
136 Fuzzy Controllers Recent Advances in Theory and Applications
Fuzzy Control of Nonlinear Systems with General Performance Criteria 19
simulation studies show that the proposed method provides a satisfactory alternative to the
existing non-linear control approaches.
Author details
Xin Wang
Oregon Institute of Technology, Department of Electrical and Renewable Energy Engineering, Klamath
Falls, Oregon, USA
Edwin E. Yaz
Marquette University, Department of Electrical and Computer Engineering, Haggerty Hall of
Engineering, Milwaukee, Wisconsin, USA
James Long
Oregon Institute of Technology, Department of Computer Systems Engineering Technology, Klamath
Falls, Oregon, USA
Tim Miller
Green Lite Motors Corporation, Portland, OR, USA
8. References
[1] Baumann W.T, Rugh W.J (1986) Feedback Control of Non-linear Systems by Extended
Linearization. IEEE Trans. Automatic Control. Vol. AC-31, No.1, pp.40-46.
[2] Basar T and Bernhard P (1995) H-innity Optimal Control and Related Minimax Design
Problems, A Dynamic Game Approach, 2nd Ed.,Birkhauser, 1995.
[3] Boyd S, Ghaoui L E, Feron E, Balakrishnan V(1994) Linear Matrix Inequalities in System
and Control Theory, SIAM Studies in Applied Mathematics, SIAM, Philadelphia.
[4] Huang Y, Lu W-M (1996) Non-linear Optimal Control: Alternatives to Hamilton-Jacobi
Equation, Proc. of 35th Conf. on Decision and Control, Kobe, Japan, pp. 3942-3947.
[5] Khalil H.K (2002) Non-linear Systems, 3rd Ed., Prentice Hall, N.J.
[6] Mohseni J, Yaz E, Olejniczak K (1998) State Dependent LMI Control of Discrete-Time
Non-linear Systems, Proc. of the 37th IEEE Conference on Decision and Control, Tampa,
FL, pp. 4626-4627.
[7] Takagi T, Sugeno M(1985) Fuzzy Identication of Systems and Its Applications to Model
and Control, IEEE Trans. Syst. Man. Cyber., Vol. 15, pp.116-132.
[8] Tanaka K, Sugeno M (1990) Stability Analysis of Fuzzy Systems Using Lyapunovs
Direct Method, Proc. NAFIPS90, pp. 133-136.
[9] Tanaka K, Ikeda T, Wang H.O (1996) Design of Fuzzy Control Systems Based on Relaxed
LMI Stability Conditions, the 35th IEEE Conference on Decision and Control, Kobe,
Vol.1, pp. 598-603.
[10] Tanaka K, Wang H.O (2001) Fuzzy Control Systems Design and Analysis, A Linear
Matrix Inequality Approach, Wiley.
[11] Van der Shaft A.J (1993) Non-linear State Space H1 control Theory, in Perspectives in
control, H. J. Trentelman and J. C. Willems, Eds. Birkhauser.
[12] Vidyasagar M (2002) Non-linear System Analysis, 2nd Ed., SIAM.
[13] Wang L.X (1994) Adaptive Fuzzy Systems and Control: Design and Stability Analysis,
Prentice Hall, Englewood Cliffs, NJ.
137 Fuzzy Control of Nonlinear Systems with General Performance Criteria
20 Fuzzy Controllers
[14] Wang H.O, Tanaka K, Grifn M (1996) An Approach to Fuzzy Control of Non-linear
Systems: Stability and Design Issues, IEEE Trans. Fuzzy Syst., Vol. 4, No. 1, pp.14-23.
[15] Wang X, Yaz E.E (2009) The State Dependent Control of Continuous-Time Non-linear
Systems with Mixed Performance Criteria, Proc. of IASTED Int. Conf. on Identi
cation Control and Applications, Honolulu, HI, pp. 98-102.
[16] Wang X, Yaz E.E (2010) Robust multi-criteria optimal fuzzy control of continuous-time
non-linear systems, Proc. of the 2010 American Control Conference, Baltimore, MD,
USA, pp. 6460-6465.
[17] Wang X, Yaz E.E, Jeong C.S (2010) Robust non-linear feedback control of discrete-time
non-linear systems with mixed performance criteria, Proc. of the 2010 American Control
Conference, Baltimore, MD, USA,pp. 6357-6362.
[18] Wang X, Yaz E.E (2010) Robust multi-criteria optimal fuzzy control of discrete-time
non-linear systems, Proc. of the 49th IEEE Conference on Decision and Control, Atlanta,
Georgia, USA, pp. 4269-4274.
[19] Wang X, Yaz E.E, Yaz Y.I (2010) Robust and resilient state dependent control of
continuous-time non-linear systems with general performance criteria, Proc. of the 49th
IEEE Conference on Decision and Control, Atlanta, Georgia, USA, pp. 603-608.
[20] Wang X, Yaz E.E, Yaz Y.I, Robust and resilient state dependent control of discrete time
non-linear systems with general performance criteria, Proc. of the 18th IFAC Congress,
Milano, Italy, pp. 10904-10909.
138 Fuzzy Controllers Recent Advances in Theory and Applications
Chapter 0
A New Method for Tuning PID-Type Fuzzy
Controllers Using Particle SwarmOptimization
S. Bouallgue, J. Haggge and M. Benrejeb
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/47139
1. Introduction
The complexity of dynamic system, especially when only qualitative knowledge about the
process is available, makes it generally difcult to elaborate an analytic model which is
sufciently precise enough for the control. Thus, it is interesting to use, for this kind of
systems, non conventional control techniques, such as fuzzy logic, in order to achieve high
performances and robustness [8, 15, 2022, 24, 33, 34]. Fuzzy logic control approach has
been widely used in many successful industrial applications which have demonstrated high
robustness and effectiveness properties.
In the literature, various Fuzzy Controller (FC) structures are proposed and extensively
studied. The particular structure given by Qiao and Mizumoto in [26], namely PID-type FC,
is especially established and improved within the practical framework in [11, 16, 31]. Such a
FC structure, which retains the characteristics similar to the conventional PID controller, can
be decomposed into the equivalent proportional, integral and derivative control components
as shown in [26]. In order to improve further the performance of the transient and steady
state responses of this kind of fuzzy controller, various strategies and methods are proposed
to tune the PID-type fuzzy controller parameters.
Indeed, Qiao and Mizumoto [26] designed a parameter adaptive PID-type FC based on a
peak observer mechanism. This self-tuning mechanism decreases the equivalent integral
control component of the fuzzy controller gradually with the system response process time.
On the other hand, Woo et al. [31] developed a method to tune the scaling factors related to
integral and derivative components of the PID-type FC structure via two empirical functions
and based on the systems error information. In [12, 16], the authors proposed a technique
that adjusts the scaling factors, corresponding to the derivative and integral components of
the PID-type FC, using a fuzzy inference mechanism. However, the major drawback of all
these PID-type FC structures is the difcult choice of their relative scaling factors. Indeed, the
fuzzy controller dynamic behaviour depends on this adequate choice. The tuning procedure
depends on the control experience and knowledge of the human operator, and it is generally
Chapter 6
2 Will-be-set-by-IN-TECH
achieved based on a classical trials-errors procedure. Up to now, there is neither clear mor
systematic method to guide such a choice. So, this tuning problem becomes more delicate
and harder as the complexity of the controlled plant increases. Hence, the proposition of a
systematic approach to tune the scaling factors of these particular PID-type FC structures is
interesting.
In this study, a new approach based on the Particle Swarm Optimization (PSO) meta-heuristic
technique is proposed for systematically tuning the scaling factors of the PID-type FC, both
with and without self-tuning mechanisms. This work can be considered as an extension of the
results given in [11, 12, 16, 26, 31]. The fuzzy control design is formulated as a constrained
optimization problem which is efciently solved based on a developed PSO algorithm.
In order to specify more robustness and performance control objectives of the proposed
PSO-tuned PID-type FC, different optimization criteria are considered and compared subject
to several various control constraints dened in the time-domain framework.
The remainder of this chapter is organized as follows. In Section 2, the proposed fuzzy
PID-type FC structures, both with and without self-tuning scaling factors mechanisms, are
presented and discussed within the discrete-time framework. Two adaptive mechanisms
for scaling factors tuning are especially adopted. The optimization-based problems of the
PID-type FC scaling factors tuning are formulated in Section 3. The developed constrained
PSO algorithm, used in solving the formulated problems, is also described. An external static
penalty technique is used to deal with optimization constraints. Theoretical conditions for
convergence algorithm and parameters choice are established, based on the stability theory of
dynamic systems. Section 4 is dedicated to apply the proposed fuzzy control approaches on
an electrical DC drive benchmark and a thermal process within an experimental real-time
framework based on an Advantech PCI-1710 multi-functions board associated with a PC
computer and MATLAB/Simulink environment. Performances on convergence properties
of the proposed PSO and the used GAO algorithm, are compared for the known Integral
Absolute Error (IAE) and the Integral Square Error (ISE) criterion cases. The real-time fuzzy
controllers are developed through the compilation and linking stage, in a form of a Dynamic
Link Library (DLL) which is, then, loaded in memory and started-up.
2. PID-type fuzzy control design
In this section, the considered PID-type FC structures are briey described within the
discrete-time framework based on [11, 12, 16, 26, 31].
2.1. Discrete-time PID-type FLC
Proposed by Qiao and Mizumoto in [26] within continuous-time formalism, this particular
fuzzy controller structure, called PID-type FC, retains the characteristics similar to the
conventional PIDcontroller. This result remains valid while using a type of FCwith triangular
and uniformly distributed membership functions for the fuzzy inputs and a crisp output, a
product-sum inference and a center of gravity defuzzication methods.
Under these conditions, the equivalent proportional, integral and derivative control
components of such a PID-type FC are given by K
e
P +K
d
D, K
e
P, and K
d
D, respectively,
as shown in [16, 26, 31]. In these expressions, P and D represent relative coefcients, K
e
, K
d
,
140 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 3
and denote the scaling factors associated to the inputs and output of the FC, as shown in
Figure 1. The proof of this computation is shown with more details in [26].
When approximating the integral and derivative terms within the discrete-time framework,
we can consider the closed-loop control structure for a discrete-time PID-type FC, as shown
in Figure 1.
Fuzzy Controller
Delay Operator
z
1
Delay operator
z
1
K
K
e
d

u
k
e
k
e
k
+
_
+
+
+
+
Figure 1. The proposed discrete-time PID-type FC structure.
As shown in [11, 16, 26, 31], the dynamic behaviour of this PID-type FC structure is strongly
dependent on the scaling factors K
e
, K
d
, and , difcult and delicate to tune.
2.2. PID-type FC with self-tuning mechanisms
In order to improve the performances of the considered PID-type FC structure, various
self-tuning mechanisms for scaling factors have been proposed in the literature. Two methods
are especially adopted in this chapter.
2.2.1. Self-tuning via Empirical Functions Tuner Method EFTM
In this self-tuning method [31], the PID-type FC integral and derivative components updating
are achieved based on scaling factors and K
d
, using the information on systems error as
follows:

k
=
0
(e
k
)
K
dk
= K
d0
(e
k
)
(1)
where
0
and K
d0
are the initial values of and K
d
, respectively, (.) and (.) are the
empirical tuner functions dened, respectively, by:
(e
k
) =
1
|e
k
| +
2
(e
k
) =
1
(1 |e
k
|) +
2
(2)
In these equations, the parameters to be tuned
1
,
2
,
1
and
2
are all positive. The empirical
function related to integral component decreases as the error decreases while the function
related to derivative factor increases. Indeed, the objective of the function is to decrease
the parameter with the change of error. However, the function has an inverse objective to
make constant the proportional effect. Hence, the system may not always keep quick reaction
against the error as demonstrated by Woo et al. in [31].
141 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
4 Will-be-set-by-IN-TECH
2.2.2. Self-tuning via Relative Rate Observer Method RROM
In this self-tuning method [12, 16], the PID-type FC integral and derivative components
updating are achieved as follows:

k
=

0
K
f

k
K
dk
= K
d0
K
f d
K
f

k
(3)
where
k
is the output of the fuzzy Relative Rate Observer (RRO) K
f
is the output scaling
factor for
k
and K
f d
is the additional parameter that affects only the derivative factor of the
FC.
The rule-base for
k
, as used by Eksin et al. [12] and Gzelkaya et al. [16], is considered for
the fuzzy RRO. This fuzzy RRO block has as inputs the absolute values of error |e
k
| and the
variable r
k
, dened subsequently, as shown in Table 1.
|e
k
|/r
k
S M F
S M M L
SM SM M L
M S SM M
L S S SM
Table 1. Fuzzy rule-base for the variable
k
.
The linguistic levels assigned to the input |e
k
| and the output variable
k
are as follows: L
(Large), M (Medium), SM (Small Medium) and S (Small). For the input variable r
k
, the
following linguistic levels are assigned: F (Fast), M (Moderate) and S (Slow).
The variable r
k
, dened in [12, 16] and called normalized acceleration, gives relative rate
information about the fastness or slowness of the system response as shown in Table 2. It is
dened as follows [16]:
r
k
=
e
k
e
k1
e

=
(e
k
)
e

(4)
where e
k
and (e
k
) are the incremental change in error and the so-called acceleration in
error given respectively by:
e
k
= e
k
e
k1
(5)
(e
k
) = e
k
e
k1
(6)
In equation (4), the variable e

is chosen as follows:
e

=
_
e
k
i f |e
k
| |e
k1
|
e
k1
i f |e
k
| < |e
k1
|
(7)
e

(e
k
) System response
Positive Positive Fast
Positive Negative Slow
Negative Positive Slow
Negative Negative Fast
Table 2. Nature of the system response depending on the variable r
k
.
142 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 5
For this RROM self-tuning approach, the uniformly distributed triangular and the
symmetrical membership functions, as shown in Figures 2, 3, 4, are assigned for the fuzzy
inputs r
k
and |e
k
|, and fuzzy output variable
k
. The view of the above fuzzy rule-base is
illustrated in Figure 5.
-1 -0. 8 -0. 6 -0. 4 -0. 2 0 0.2 0.4 0.6 0.8 1
0
0.2
0.4
0.6
0.8
1
r
k
D
e
g
r
e
e

o
f

m
e
m
b
e
r
s
h
i
p


S
M F
Fuzzy input
Figure 2. Membership functions for r
k
.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0.2
0.4
0.6
0.8
1
e
k
D
e
g
r
e
e

o
f

m
e
m
b
e
r
s
h
i
p

S M L SM
Fuzzy input | |
Figure 3. Membership functions for |e
k
|.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0.2
0.4
0.6
0.8
1

k
D
e
g
r
e
e

o
f

m
e
m
b
e
r
s
h
i
p

S M L SM
Fuzzy output
Figure 4. Membership functions for
k
.
143 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
6 Will-be-set-by-IN-TECH
-1
-0. 5
0
0.5
1
0
0.2
0.4
0.6
0.8
1
0.2
0.4
0.6
0.8
r
k
e
k

k
Figure 5. View of the fuzzy rule-base for
k
.
3. The proposed PSO-based approach
In this section, the problem of scaling factors tuning, for all dened PID-type FC structures,
is formulated as a constrained optimization problem which is solved using the proposed
PSO-based approach.
3.1. PID-type FC tuning problem formulation
The choice of the adequate values for the scaling factors of each PID-type FC structure is often
done by a trials-errors hard procedure. This tuning problem becomes difcult and delicate
without a systematic design method. To deal with these difculties, the optimization of these
scaling factors is proposed like a promising solution. This tuning problem can be formulated
as the following constrained optimization problem:

minimize
xxxD
f (xxx)
subjectto
g
l
(xxx) 0; l = 1, . . . , n
con
(8)
where f : R
m
Rthe cost function, D = {xxx D
m
; xxx
min
xxx xxx
max
} the initial search space,
which is supposed containing the desired design parameters, and g
l
: R
m
R the problems
constraints.
The optimization-based tuning problem consists in nding the optimal decision variables
xxx

=
_
x

1
, x

2
, . . . , x

m
_
T
, representing the scaling factors of a given PID-type FC structure,
which minimize the dened cost function, chosen as the ISE and IAE performance criteria.
These cost functions are minimized, using the proposed constrained PSO algorithm, under
various time-domain control constraints such as overshoot D, steady state error E
ss
, rise time
t
r
and settling time t
s
of the systems step response, as shown in the equations (9), (10) and (11).
Hence, in the case of the PID-type FC structure without self-tuning mechanisms, the scaling
factors to be optimized are K
e
, K
d
, and . The formulated optimization problem is dened
as follows:

minimize
xxx=(K
e
,K
d
,,)
T
R
4
+
f (xxx)
subjectto
D D
max
; t
s
t
max
s
; t
r
t
max
r
; E
ss
E
max
ss
(9)
144 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 7
where D
max
, E
max
ss
, t
max
r
and t
max
s
are the specied overshoot, steady state, rise and settling
times respectively, that constraint the step response of the PSO-tuned PID-type FC controlled
system, and can dene some time-domain templates.
In the case of the PID-type FC structure with the EFTM self-tuning mechanism, the scaling
factors to be optimized are
1
,
2
,
1
and
2
. The formulated optimization problemis dened
as follows:

minimize
xxx=(
1
,
2
,
1
,
2
)
T
R
4
+
f (xxx)
subjectto
D D
max
; t
s
t
max
s
; t
r
t
max
r
; E
ss
E
max
ss
(10)
For the PID-type FC structure with the RROM self-tuning mechanism, the scaling factors to
be optimized are K
f
and K
f d
. The formulated optimization problem is dened as follows:

minimiser
xxx=(K
f
,K
f d)
T
R
2
+
f (xxx)
subjectto
D D
max
; t
s
t
max
s
; t
r
t
max
r
; E
ss
E
max
ss
(11)
3.2. Particle Swarm Optimization technique
In this study, the proposedPSOapproach is presented and a constrained PSOalgorithmis also
developed. The convergence conditions of such an algorithm are analyzed and established.
3.2.1. Overview
The PSO technique is an evolutionary computation method developed by Kennedy and
Eberhart [9]. This recent meta-heuristic technique is inspiredby the swarming or collaborative
behaviour of biological populations. The cooperation and the exchange of information
between population individuals allow solving various complex optimization problems [10,
25, 27, 28, 30].
Without any regularity on the cost function to be optimized, the recourse to this stochastic
and global optimization technique is justied by the empirical evidence of its superiority in
solving a variety of non-linear, non-convex and non-smooth problems. In comparison with
other meta-heuristics, this optimization technique is a simple concept, easy to implement, and
a computationally efcient algorithm[10, 27, 30]. The convergence and parameters selection of
the PSO algorithmare proved using several advanced theoretical analysis proposed by Ruben
and Kamran in [27] and Van den Bergh in [30]. Its stochastic behaviour allows overcoming
the local minima problem.
Particle swarmoptimisation has been enormously successful in several and various industrial
domains [18, 19]. It has been used across a wide range of engineering applications. These
applications can be summarized around domains of robotics, image and signal processing,
electronic circuits design, communication networks, but more especially the domain of plant
control design, as shown in [26].
145 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
8 Will-be-set-by-IN-TECH
3.2.2. Basic PSO algorithm
The basic PSO algorithm uses a swarm consisting of n
p
particles (i.e. xxx
1
, xxx
2
, . . . , xxx
n
p
),
randomly distributed in the considered initial search space, to nd an optimal solution
xxx

= arg min f (xxx) R

m
of a generic optimization problem (8). Each particle, that
represents a potential solution, is characterised by a position and a velocity given by xxx
i
k
:=
_
x
i,1
k
, x
i,2
k
, . . . , x
i,m
k
_
T
and vvv
i
k
:=
_
v
i,1
k
, v
i,2
k
, . . . , v
i,m
k
_
T
where (i, k)
__
1, n
p

[[1, k
max
]].
At each algorithm iteration, the i
th
particle position, xxx
i
R
m
, evolves based on the following
update rules:
xxx
i
k+1
= xxx
i
k
+vvv
i
k+1
(12)
vvv
i
k+1
= w
k+1
vvv
i
k
+ c
1
r
i
1,k
_
ppp
i
k
x
i
k
_
+ c
2
r
i
2,k
_
ppp
g
k
x
i
k
_
(13)
where
w
k+1
: the inertia factor,
c
1
, c
2
: the cognitive and the social scaling factors respectively,
r
i
1,k
, r
i
2,k
: random numbers uniformly distributed in the interval [[0, 1]],
ppp
i
k
: the best previously obtained position of the i
th
particle,
ppp
g
k
: the best obtained position in the entire swarm at the current iteration k.
Hence, the principle of a particle displacement in the swarm is graphically shown in the
Figure 6, for a two dimensional design space.
........ .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
x
x
p
p
v
v
v
p x
k
i
k
i
k
i
k
g
k
i
k
i
i i
k
i
k
i
w c
+
+
_
(
1
1
1
) k
r
k , 2
i
r
k , 1
c
2
p x
k
g
k
i
_
( )
Figure 6. Particle position and velocity updates.
In order to improve the exploration and exploitation capacities of the proposed PSO
algorithm, we choose for the inertia factor a linear evolution with respect to the algorithm
iteration as given by Shi and Eberhart in [28]:
w
k+1
= w
max

_
w
max
w
min
k
max
_
k (14)
where w
max
= 0.9 and w
min
= 0.4 represent the maximumand minimuminertia factor values,
respectively, k
max
is the maximum iteration number.
146 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 9
Similarly to other meta-heuristic methods, the PSO algorithm is originally formulated as an
unconstrained optimizer. Several techniques have been proposed to deal with constraints.
One useful approach is by augmenting the cost function of problem (8) with penalties
proportional to the degree of constraint infeasibility. In this paper, the following external
static penalty technique is used:
(xxx) = f (xxx) +
n
con

l=1

l
max
_
0, g
l
(xxx)
2
_
(15)
where
l
is a prescribed scaling penalty parameters and n
con
is the number of problem
constraints g
l
(xxx).
Finally, the basic proposed PSO algorithm can be summarized by the following steps:
1. Dene all PSO algorithm parameters such as swarm size n
p
, maximum and minimum
inertia factor values, cognitive c
1
and social c
2
scaling factors, etc.
2. Initialize the n
p
particles with randomly chosen positions xxx
i
0
and velocities vvv
i
0
in the search
space D. Evaluate the initial population and determine ppp
i
0
and ppp
g
0
.
3. Increment the iteration number k. For each particle apply the update equations (12)
and (13), and evaluate the corresponding tness values
i
k
=
_
xxx
i
k
_
:
if
i
k
pbest
i
k
then pbest
i
k
=
i
k
and ppp
i
k
= xxx
i
k
;
if
i
k
gbest
k
then gbest
k
=
i
k
and ppp
g
k
= xxx
i
k
;
where pbest
i
k
and gbest
k
represent the best previously tness of the i
th
particle and the
entire swarm, respectively.
4. If the termination criterion is satised, the algorithm terminates with the solution xxx

=
arg min
xxx
i
k
_
f
_
xxx
i
k
_
, i, k
_
. Otherwise, go to step 3.
3.2.3. The convergence of PSO algorithm analysis
In this part, the proposed PSO algorithm is analysed based on results in [27, 30]. Theoretical
conditions for convergence algorithm and parameters choice are established.
Let us replace the velocity update equation (13) into the position update equation (12) to get
the following expression:
xxx
i
k+1
=
_
1 c
1
r
i
1,k
c
2
r
i
2,k
_
xxx
i
k
+ wvvv
i
k
+ c
1
r
i
1,k
ppp
i
k
+ c
2
r
i
2,k
ppp
g
k
(16)
A similar re-arrangement of the velocity term (13) leads to:
vvv
i
k+1
=
_
c
1
r
i
1,k
+ c
2
r
i
2,k
_
xxx
i
k
+ wvvv
i
k
+ c
1
r
i
1,k
ppp
i
k
+ c
2
r
i
2,k
ppp
g
k
(17)
The obtained equations (16) and (17) can be combined and written in matrix form as:
_
xxx
i
k+1
vvv
i
k+1
_
=

1
_
c
1
r
i
1,k
+ c
2
r
i
2,k
_
w

_
c
1
r
i
1,k
+ c
2
r
i
2,k
_
w

_
xxx
i
k
vvv
i
k
_
+
_
c
1
r
i
1,k
c
2
r
i
2,k
c
1
r
i
1,k
c
2
r
i
2,k
_
_
ppp
i
k
ppp
g
k
_
(18)
147 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
10 Will-be-set-by-IN-TECH
This above expression can be considered as a state-space representation of a discrete-time
dynamic linear system, given by:
y
k+1
= M y
k
+N u
k
(19)
where y
k
is the state vector, u
k
the external input system, M and N the dynamic and input
matrices respectively, dened as:
y
k
=
_
xxx
i
k
vvv
i
k
_
; u
k
=
_
ppp
i
k
ppp
g
k
_
; M =

1
_
c
1
r
i
1,k
+ c
2
r
i
2,k
_
w

_
c
1
r
i
1,k
+ c
2
r
i
2,k
_
w

; N =
_
c
1
r
i
1,k
c
2
r
i
2,k
c
1
r
i
1,k
c
2
r
i
2,k
_
(20)
For a given particle, the convergent behaviour can be maintained while assuming that the
external input is constant, as there is no external excitation in the dynamic system. In such
a case, as the iterations go to innity the updated positions and velocities become constants
from the k
th
to the (k +1)
th
iteration, given the following equilibrium state:
y
k+1
y
k
=

_
c
1
r
i
1,k
+ c
2
r
i
2,k
_
w

_
c
1
r
i
1,k
+ c
2
r
i
2,k
_
w1

_
xxx
i
k
vvv
i
k
_
+
_
c
1
r
i
1,k
c
2
r
i
2,k
c
1
r
i
1,k
c
2
r
i
2,k
_
_
ppp
i
k
ppp
g
k
_
=
_
0
0
_
(21)
which is true only when:
xxx
i
k
= ppp
i
k
= ppp
g
k
,
vvv
i
k
= 0
(22)
Therefore, we obtain an equilibrium point, for which all particles tend to converge as
algorithm iteration progresses, given by:
y
eq
=
_
ppp
g
k
, 0
_
T
(23)
So, the dynamic behaviour of the i
th
particle can be analysed using the eigenvalues derived
from the dynamic matrix formulation (19) and (20), solutions of the following characteristic
polynomial:

_
1 + wc
1
r
i
1,k
c
2
r
i
2,k
_
+ w = 0 (24)
The following necessary and sufcient conditions for stability of the considered discrete-time
dynamic system (20) are obtained while applying the classical Jury criterion:
|w| < 1
c
1
r
i
1,k
+ c
2
r
i
2,k
> 0
w +1
c
1
r
i
1,k
+c
2
r
i
2,k
2
> 0
(25)
Knowing that r
i
1,k
, r
i
2,k
[[0, 1]], the above stability conditions are equivalents to the following
set of parameter selection heuristics which guarantee convergence for the PSO algorithm:
0 < c
1
+ c
2
< 4
c
1
+c
2
2
1 < w < 1
(26)
While these heuristics provide useful selection parameter bounds, an analysis of the effect of
the different parameter settings is achieved and veried by some numerical simulations to
determine the effect of such parameters in the PSO algorithm convergence performances.
148 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 11
In order to illustrate the efciency of the proposed PSO algorithm in the resolution of
problems (9), (10) and (11), several comparisons with the Genetic Algorithms Optimization
GAO-based method [14, 29] are considered. The next section is dedicated to the application of
the proposed PSO-tuned PID-FC approaches to an electrical DC drive and a thermal process
within a developed real-time framework.
4. Real-time control approach implementation
In this section, all designed PSO-tuned PID-type FC structures are applied to two different
systems such as an electrical DC drive and a thermal PT-326 Process Trainer benchmarks.
Real-time implementations and experimental results of these control laws are presented and
discussed.
4.1. Control of an electrical DC drive benchmark
4.1.1. Plant model description
The considered benchmark is a 250 watts electrical DC drive, as shown in Figure 15. The
machines speed rotation is 3000 rpmat 180 volts DC armature voltage. The motor is supplied
by an AC-DC power converter. The developed real-time application acquires input data
(speed of the DC drive) and generates control signal for thyristors of AC-DC power converter
(PWM signal). This is achieved using a data acquisition and control system based on a PC
computer and a multi-functions data acquisition PCI-1710 board which is compatible with
MATLAB/Simulink [1, 17].
The considered electrical DC drive can be described by the following model that is used in the
design setup:
G (s) =
k
m
(1 +
m
s) (1 +
e
s)
(27)
The models parameters are obtained by an experimental identication procedure and they are
summarized in Table 3 with their associated uncertainty bounds. Also, this model is sampled
with 10 ms sampling time for simulation and experimental setups.
Parameters Nominal values Uncertainty bounds
k
m
0.05 75 %

m
300 ms 75 %

e
14 ms 75 %
Table 3. Identied DC drive model parameters.
4.1.2. Simulation results
For all proposed PSO-tuned PID-type FC structures, product-sum inference and center of
gravity defuzzication methods are adopted for the FC block. Uniformly distributed and
symmetrical membership functions, are assigned for the fuzzy input and output variables.
The associated fuzzy rule-base is given in Table 4.
149 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
12 Will-be-set-by-IN-TECH
e
k
/ e
k
N Z P
N NB N Z
Z N Z P
P Z P PB
Table 4. Fuzzy rule-base for the output u
f z
.
The linguistic levels assigned to the input variables e
k
and e
k
, and the output variable u
f z
are
given as follows: N (Negative), Z (Zero), P (Positive), N (Negative), NB (Negative Big) and PB
(Positive Big). The view of this rule-base is illustrated in Figure 7.
-1
-0. 5
0
0.5
1
-1
-0. 5
0
0.5
1
-0. 5
0
0.5
e
k
e
k
u
f
z
Figure 7. View of fuzzy rule-base for the fuzzy output u
f z
.
The swarm size algorithms choice is generally a problem-dependent in PSO framework.
However, Eberhart and Shi [10] as well as Poli et al. [25] show that this parameter is often
set empirically in relation to the dimensionality and perceived difculty of a considered
optimization problem. They suggest that swarm size values in the range 20-50 are quite
common. For this purpose, we have tested the proposed PSO algorithm with different values
in this range for the case of PID-type FC structure without self-tuning mechanisms. Globally,
all the found results are close to each other. But, best values of the tness are obtained while
using a swarm size equal to 30. Henceforth, this value will be adopted for our following
works.
In the PSO framework, it is necessary to run the algorithm several times in order to get
some statistical data on the quality of results and so to validate the proposed approach. We
run the proposed algorithm 20 times and feasible solutions are found in 98 % of trials and
within an acceptable CPU computation time for the IAE and ISE criterion cases. The obtained
optimization results are summarized in Tables 5, 6 and 7. Besides, the fact that the algorithms
convergence always takes place in the same region of the design space, whatever is the
initial population, indicates that the algorithm succeeds in nding a region of the interesting
research space to explore. The performances comparison of PSO- and GAO-based approaches
is achieved in the same conditions.
Cost function Algorithm Best Mean Worst St. dev.
ISE PSO 0.0193 0.0304 0.0511 0.018
ISE GAO 0.1200 0.1780 0.2410 0.050
IAE PSO 0.0162 0.0261 0.0497 0.016
IAE GAO 0.1892 0.2835 0.3227 0.066
Table 5. Optimization results from 20 trials of problem (9).
150 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 13
Cost function Algorithm Best Mean Worst St. dev.
ISE PSO 0.0660 0.0765 0.1030 0.015
ISE GAO 0.0820 0.0912 0.1330 0.012
IAE PSO 0.0659 0.0838 0.0946 0.014
IAE GAO 0.0718 0.0814 0.0973 0.013
Table 6. Optimization results from 20 trials of problem (10).
Cost function Algorithm Best Mean Worst St. dev.
ISE PSO 0.0559 0.0792 0.0840 0.013
ISE GAO 0.0822 0.0936 0.1120 0.015
IAE PSO 0.0673 0.0861 0.0993 0.016
IAE GAO 0.0855 0.0905 0.1009 0.008
Table 7. Optimization results from 20 trials of problem (11).
Indeed, the population size, used in the GAO algorithm, is set as 30 individuals and the
maximum generation number is 50. However, the GA parameters, used for MATLAB
simulations, are chosen as the Stochastic Uniform selection and the Gaussian mutation
methods. The Elite Count is set as 2 and the Crossover Fraction as 0.8. The algorithm stops
when the number of generations reaches the specied value for the maximum generation.
According to the statistical analysis of Tables 5,6 and 7, we can conclude that the proposed
PSO-based approach produces better results in comparison with the standard GAO-based
one. Also, while using a Pentium IV, 1.73 GHz and MATLAB 7.7.0, the CPU computation
times are about 358 and 364 seconds for ISE and IAE criteria, respectively, for the considered
PID-type FC without self-tuning mechanisms structure.
On the other hand, performances on convergence properties of the proposed PSO and the
used GAO algorithm, in term of iterations numbers required to nd the best solution,
are compared for the IAE criterion case, as shown in Figures 8 and 9. While using the
proposed PSO-based method, we succeed to obtain the optimal solution within only about
28 iterations. However, the GAO-based method nds the same result after 40 iterations.
All these observations can show the superiority of the proposed PSO-based method in
comparison with the GAO-based one. Indeed, the quality of the obtained optimal solution,
the fastness convergence as well as the simple software implementation is better than those of
the GAO-based approach.
In a typical optimization procedure, the scaling parameters
l
, given in equation (15), will be
linearly increased at each iteration step so constraints are gradually enforced. Generally, the
quality of the solution will directly depend on the value of the specied scaling parameters.
In this paper and in order to make the proposed approach simple, great and constant scaling
penalty parameters, equal to 10
3
, are used for simulations. Indeed, simulation results show
that with a great values of
l
, the control system performances are weakly degraded and the
effects on the tuning parameters are less meaningful. The PSOalgorithmconvergence is faster
than the case with linearly variable scaling parameters.
The robustness of the proposed PSO algorithm convergence, under variation of the cognitive,
social and inertia factor parameters, is analysed based on numerical simulations as shown
151 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
14 Will-be-set-by-IN-TECH
0 10 20 30 40 50
0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
Iteration
C
o
s
t

f
u
n
c
t
i
o
n

v
a
l
u
e
IAE*=0.0162
Figure 8. Convergence properties of the proposed PSO algorithm: IAE criterion case.
0 10 20 30 40 50
0.15
0. 2
0.25
0.3
0.35
0.4
0.45
Generation
C
o
s
t

f
u
n
c
t
i
o
n

v
a
l
u
e


IAE*=0.1892
Figure 9. Convergence properties of the standard GAO algorithm: IAE criterion case.
in Figure 10 and Figure 11. The PSO algorithms convergence is guaranteed within the
established domain given by the equation (26).
0 10 20 30 40 50
-0. 1
0
0.1
0.2
0.3
0.4
0.5
0.6
Iteration
C
o
s
t

f
u
n
c
t
i
o
n

v
a
l
u
e


c
1
=0; c
2
=3
c
1
=3; c
2
=0
c
1
=1.75; c
2
=0.8
c
1
=0.5; c
2
=2.5
c
1
=2; c
2
=2
c
1
=1.19; c
2
=1.19
(a)
Figure 10. Robustness of the proposed PSO algorithm under variations of the cognitive and social
parameters: IAE criterion case.
152 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 15
0 10 20 30 40 50
-0.02
0
0.02
0.04
0.06
0.08
0.1
Iteration
C
o
s
t

f
u
n
c
t
i
o
n

v
a
l
u
e


w
max
=0.75, w
min
=0.4
w
max
=0.9, w
min
=0.1
w
max
=0.8, w
min
=0.2
w
max
=0.6, w
min
=0.3
w
max
=0.95, w
min
=0.2
w
max
=0.75, w
min
=0.25
Figure 11. Robustness of the proposed PSO algorithm under variations of the inertia factor: IAE
criterion case.
The robust stability of the proposed PSO-tuned PID-type FC approach is analysed while
considering external disturbances and model uncertainties. According to uncertain bounds
on nominal plant parameters, given in Table 1, we are going to consider the following family
of continuous-time transfer functions supposed including the real studied plant:
G =
_

G(s) =
k
m
(1 +
m
s) (1 +
e
s)
; k
m

_
k
min
m
, k
max
m
_
,
e

_

min
e
,
max
e
_
,
m

_

min
m
,
max
m
_
_
(28)
Figure 12 shows the step responses of a family of 5 random generated closed-loop uncertain
models. The stability robustness of the uncertain plants, under the above considered
uncertainty types, is guaranteed for all designed PID-type FC structures.
0 0.5 1 1.5 2 2.5 3
0
0.5
1
1.5
Time (sec)
D
C

d
r
i
v
e

c
o
n
t
r
o
l
l
e
d

s
p
e
e
d

(
1
0
0
0

r
p
m
)


nominal plant
uncertain plant
(IAE criterion case)
0 0.5 1 1.5 2 2.5 3
0
0.5
1
1.5
Time (sec)
D
C

d
r
i
v
e

c
o
n
t
r
o
l
l
e
d

s
p
e
e
d

(
1
0
0
0

r
p
m
)


nominal plant
uncertain plant
(ISE criterion case)
Figure 12. Stability robustness of the PSO-tuned PID-type fuzzy controlled system under model
parameters uncertainties and external disturbances.
Finally, the time-domain performances of all proposed PID-type FC structures, are compared
for the PSO- and GAO-based design cases as shown in Figure 13.
Besides, Table 8 shows the superiorities of the self-tuning EFTM and RROM PID-type FC
structures in relation to the one without self-tuning mechanisms as veried in [16]. Remenber
that the considered time-domain constraints for the PID-type FC tuning problems (9), 10)
153 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
16 Will-be-set-by-IN-TECH
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Time (sec)
D
C

d
r
i
v
e

c
o
n
t
r
o
l
l
e
d

s
p
e
e
d

(
1
0
0
0

r
p
m
)


PSO-based proposed approach
GAO-based approach
GAO:
e
=1.4720;
d
=1.9879; =1.9844; =5.2642; =0.2933
PSO:
e
=1.8240;
d
=10.7780; =0.4747; =32.8991; =0.0957
PSO-tuned PID-type FC
without self-tuning mechanisms
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Time (sec)
D
C

d
r
i
v
e

c
o
n
t
r
o
l
l
e
d

s
p
e
e
d

(
1
0
0
0

r
p
m
)


Proposed PSO-based approach
GAO-based approach
GAO:
f
=0.6980;
fd
=5.2936; =0.0989
PSO:
f
=0.5284;
fd
=8.3131; =0.0673
PSO-tuned PID-type FC
with EFTM self-tuning mechanism
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Time (sec)
D
C

d
r
i
v
e

c
o
n
t
r
o
l
l
e
d

s
p
e
e
d

(
1
0
0
0

r
p
m
)


Proposed PSO-based approach
GAO-based approach
GAO:

1
=1.1327;

2
=3.6938;

1
=1.8558;

2
=0.3326; =0.0718
PSO:

1
=1.0329;

2
=3.9495;

1
=2.4156;

2
=0.4503; =0.0659
PSO-tuned PID-type FC
with RROM self-tuning mechanism
Figure 13. Time-domain performances comparison of all designed PSO- and GAO-tuned PID-type FC
structures: IAE criterion case.
and (11) problems, dened in terms of overshoot, steady state, rise and settling times, have
been specied as D
max
= 20%, E
max
ss
= 0, t
max
r
= 0.25 sec and t
max
s
= 0.75 sec.
PSO-tuned PID-type FC structure D(%) t
r
(sec) t
s
(sec) E
ss
CPU computation time (sec)
without self-tuning mechanisms 17.5 0.23 0.49 0 364
with EFTM self-tuning mechanism 15 0.21 0.64 0 370
with RROM self-tuning mechanism 7 0.20 0.68 0 392
Table 8. Performances of the PSO-tuned PID-type FC structures: IAE criterion case.
4.1.3. Experimental setup and results
In order to illustrate the efciency of the proposed PSO-tuned fuzzy control structures within
a real-time framework, the example of the PID-type FC without self-tuning mechanism is
considered. The same principle of implementation remains valid for the other PID-type FC
structures.
The controlled process is constituted by the single-phase AC-DC power converter and the
independent excitation DC motor. A schematic diagram of the experimental setup prepared
for testing of the designed controller is shown in Figure 14. The developed experimental
154 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 17
benchmark is given by Figure 15. The designed real-time application acquires input data
(speed of the DC drive) and generates control signals for the AC-DC power converter through
a thyristors gate drive circuit. This is achieved using a data acquisition and control system
based on PC and a multi-function data acquisition PCI-1710 board with 12-bit resolution of
A/D converter and up to 100 KHz sampling rate. A thyristors gate drive circuit, based on
a multivibrator, is used to generate a triggering burst of high-frequency impulses. A pulse
transformer is used to assure the galvanic insulation between the control and power circuits.
The acquired speed measure, obtained fromtachometer sensor, must be adapted to be applied
to the used multi-function PCI-1710 board. The complete electronic circuit diagram of the
designed control system is given in [1, 17].
Load
D1 D2
Th1 Th2
DC
Tachometer
D3
motor
.
DT

v
Interfacing
of AC-DC power converter control
and galvanic insulation
Interfacing
of acquirement and adaptation
of speed DC motor
Gate impulses

.
.
~
vs
is
Data Acquisition
and Control System
Thyristors
gate drive circuit
Figure 14. The proposed experimental setup schematic.
Figure 15. The developed experimental DC drive benchmark.
The multi-function data acquisition PCI-1710 board allows achieving measurement and
controlling functions. This target is used to create a real-time application to let the
implemented controller system run while synchronized to a real-time clock. The model of
the plant was removed from the simulation model, and instead of it, the input device driver
(Analog Input) and the output device driver (Analog Output) were introduced as shown in
Figure 17. These device drivers close the feedback loop when moving from simulations to
experiments. Device drivers blocks include procedures to access the inputs-outputs board.
The real-time controller is developed through the compilation and linking stage, in a form of
a Dynamic Link Library (DLL) which is then loaded in memory and started-up
The practical implementation of the PSO-tuned PID-type FC approach leads to the
experimental results of Figure 17 and Figure 18. The obtained results are satisfactory for
a simple, systematic and non-conventional control approach and point out the controllers
viability and validate the proposed control approach. The measured speed tracking error
of the controlled DC drive is very small (less than 10 % of set point) showing the high
performances of the proposed control especially in terms of tracking. On the other hand,
155 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
18 Will-be-set-by-IN-TECH
ek
ek
Trajectory planning
block
Controlled system
output
Control signal
Tracking error
Delay operator
z
1
Delay operator
z
1
S-Function Builder
traj 1_c t yc
Ke
Kd
Fuzzy Controller Clock

Analog Output
Advantech
PCI -1710 [auto ]
Analog
Output
Analog Input
Advantech
PCI -1710 [auto ]
Analog
Input

Figure 16. PCI-1710 board based real-time implementation of the proposed PSO-tuned PID-type FC
structure.
Figure 18 shows the robustness of the proposed PSO-tuned PID-type FC in rejection of an
external load disturbance applied on the controlled system. The dynamic of the disturbance
rejection is fast and guaranteed.
0 10 20 30 40 50 60 70
0. 5
0
0.5
1
1.5
2
2.5
3
Acquisition time (sec)
C
o
n
t
r
o
l
l
e
d

s
p
e
e
d

v
a
r
i
a
t
i
o
n

(
1
0
0
0

r
p
m
)
0 10 20 30 40 50 60 70
-0. 5
-0. 4
-0. 3
-0. 2
-0. 1
0
0.1
0.2
0.3
0.4
0.5
Acquisition time (sec)
S
p
e
e
d

t
r
a
c
k
i
n
g

e
r
r
o
r

(
1
0
0
0

r
p
m
)
Figure 17. Experimental results of PSO-tuned PID-type FC implementation: fuzzy controller tracking
performances.
4.2. Control of a thermal process benchmark
4.2.1. Plant model description
The thermal process to be controlled, shown in the photography of Figure 20, is based on
a known PT-326 Process trainer [13], initially developed with an analog control system and
modied in order to be digitally controlled. To power the heating resistor, a single-phase
AC-AC converter, is developed [7].
In this prototype, the air drawn from atmosphere by a centrifugal blower is injected, through
a heating element, in a polypropylene pipe, and rejected in the atmosphere. The amount of air
156 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 19
0 10 20 30 40 50 60 70
0
1
2
3
C
o
n
t
r
o
l
l
e
d

s
p
e
e
d

v
a
r
i
a
t
i
o
n

(
1
0
0
0

r
p
m
)
External load disturbance rejection
Acquisition time (sec)
0 10 20 30 40 50 60 70
-0.5
-0.25
0
0.25
0.5
Acquisition time (sec)
S
p
e
e
d

t
r
a
c
k
i
n
g

e
r
r
o
r

(
1
0
0
0

r
p
m
)
External load disturbance rejection
Figure 18. Experimental results of PSO-tuned PID-type FC implementation: fuzzy controller robustness
under external load disturbance.
owing in the pipe can be adjusted by the mean of an inlet throttle attached to the blower. The
process consists of heating the air owing in the pipe to the desired temperature level. The
digital control systemgenerates a 40Wsignal which determines the amount of electrical power
supplied to heating resistor made of 10K/7W power resistors. According to these settings,
the experimental trials showthat the controlled air temperature can be varied up to 20

C from
the ambient temperature. The assigned control objective is to regulate the temperature of the
air at a desired level, with high tracking performance and under internal disturbances, like
model parameters variation, and output disturbances. The temperature sensor can be placed
at three different locations on the path of the air ow. A variation in the temperature makes a
voltage variation at the sensors output. The amount of air trough the pipe, adjusted by setting
the opening of the throttle, can also be used to generate an output disturbance, in order to test
the efciency of the proposed control system.
The controlled system input is the voltage applied to the AC-AC power electronic circuit
feeding the heating resistor, and the output is the air ow temperature in the pipe, expressed
by a 50 mV/

C voltage, obtained after amplication of the LM35 temperature sensors output

signal. As shown in [23, 32], this process can be characterized as a non-linear system with
a pure time delay. The pure time delay depends on the position of the temperature sensor
element inserted into the air stream at any one of the three positions along the tube. When
the temperature in the air volume inside the tube is assumed uniform a linear model can be
obtained. To identify a numerical model of the consideredplant, some experimental trials lead
to consider the following transfer function, between the heater input voltage and the sensors
output voltage:
G (s) =
k
m
1 +s
e

d
s
(29)
where k
m
is the DC gain system, is the time constant system, and
d
is the time delay system.
This obtained plant model is assumed to be the nominal one and will be adopted in PSO-tuned
PID-type FC synthesis step. These models parameters are obtained by an experimental
identication procedure and they are summarized in Table 9 with their associated uncertainty
bounds. Also, this model is sampledwith 2 sec sampling time for simulation and experimental
setups.
157 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
20 Will-be-set-by-IN-TECH
Parameters Nominal values Uncertainty bounds
k
m
20 50 %
65 sec 50 %

d
1 sec 50 %
Table 9. Identied Thermal Process model parameters.
For this PSO-tuned PID-type fuzzy control example, we represent only the obtained
experimental results. For the numerical simulations step, both IAE and ISE criterion, used
for the electrical DC drive control, are investigated for this thermal process example. Same
problems (9), (10) and (11) are considered and resolved by the developed constrained PSO
algorithm.
4.2.2. Experimental setup and results
The developed real-time application acquires air temperature measure and generates control
signals for the triac of AC-AC power converter through a gate drive circuit, as shown in
Figure 19. This is achieved using a control system based on PC and the used multi-function
data acquisition PCI-1710 board. A triac gate drive circuit is used to generate a Pulse Width
Modulation (PWM) control signal synchronized with the zero-crossing of the AC voltage. The
acquired air temperature measure is scaled before being applied to the used multi-function
PCI-1710 board used to create a real-time application to let the implemented controller system
run while synchronized to a real-time clock. This leads to experimental results shown in
Figure 21 and Figure 22.
160
90
Air Output
Disturbance PT-326 Process Trainer
Interfacing
of AC-AC power converter control
and galvanic insulation
Interfacing
of acquirement and adaptation
of Air temperature
Gate impulses

Data Acquisition
and Control System
Triac
gate drive circuit
Heater
Sensor
Figure 19. The proposed thermal process experimental setup schematic.
Figure 20. Developed experimental benchmark of the PT-326 Process Trainer.
158 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 21
As shown in Figure 21 and Figure 22, the controlled air temperature of the considered thermal
process tracks the desired trajectory with high performance in terms of response speed and
precision in the two considered cases. The robustness of the proposed control strategy in term
of output static disturbance rejection, which caused by the throttle opening, is improved.
0 500 1000 1500
20
25
30
35
40
45
Acquisition time (sec)
T
h
e
r
m
a
l

p
r
o
c
e
s
s

t
e
m
p
e
r
a
t
u
r
e

[

C
]
External output disturbance rejection
controlled temperature
desired temperature
0 500 1000 1500
-5
0
5
10
15
20
Acquisition time (sec)
T
r
a
c
k
i
n
g

e
r
r
o
r

t
e
m
p
e
r
a
t
u
r
e

[

C
]
External output disturbance rejection
Figure 21. Experimental result of PSO-tuned fuzzy controlled PT-326 Process Trainer: IAE criterion case.
0 500 1000 1500
20
25
30
35
40
45
Acquisition time (sec)
T
h
e
r
m
a
l

p
r
o
c
e
s
s

t
e
m
p
e
r
a
t
u
r
e

[

C
]
External ouput disturbance rejection
controlled temperature
desired temperature
0 500 1000 1500
5
0
5
10
15
20
Acquisition time (sec)
T
r
a
c
k
i
n
g

e
r
r
o
r

t
e
m
p
e
r
a
t
u
r
e

[

C
]
External output disturbance rejection
Figure 22. Experimental result of PSO-tuned fuzzy controlled PT-326 Process Trainer: ISE criterion case.
5. Conclusion
In this study, a newmethod for tuning PID-type FC structures, using a constrained PSO-based
technique, is proposed and successfully applied to an electrical DC drive and thermal process
within a real-time framework. This efcient tool leads to a robust and systematic fuzzy control
design approach. The performances comparison, with the standard GAO-based method,
shows the efciency and superiority of the proposed PSO-based approach in terms of the
obtained solution qualities, the convergence speed and the simple software implementation
of its algorithm.
The practical implementation of the PSO-tuned PID-type FC approach, for the considered
electrical DC drive and the thermal PT-326 Process Trainer benchmarks, leads to several
159 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
22 Will-be-set-by-IN-TECH
satisfactory experimental results showing the high performances of the proposed control
especially in terms of tracking and robustness.
The PSO-tuned PID-type FC structures robustness, under external inuences such as the
output static disturbances and parametric uncertainties, are proven. The control design
methodology is systematic, practical and simple without need to exact analytic plant model
description. The obtained simulation and experimental results show the efciency in terms of
performance and robustness of the proposed fuzzy control approach which can be applied in
industrial motor control eld.
Author details
S. Bouallgue
Higher Insitute of Industrial Systems of Gabes (ISSIG), Salaheddine Elayoubi Street, 6032 Gabes, g
Tunisia
J. Haggge and M. Benrejeb
National Engineering School of Tunis (ENIT), BP 37, le Belvdre, 1002 Tunis, Tunisia
All authors are with the Research Laboratory in Automatic Control (LA.R.A) of ENIT.
6. References
[1] Bouallgue, S., Haggge, J., Benrejeb, M., (2012). On a robust real-time H

controller
design for an electrical drive, International Journal of Modelling, Identication and
Control, 15 (2), pp. 89-96.
[2] Bouallgue, S., Haggge, J., Ayadi, M., Benrejeb, M., (2012). PID-type fuzzy logic
controller tuning based on particle swarm optimization, Engineering Applications of
Articial Intelligence, 25 (3), pp. 484-493.
[3] Bouallgue, S., (2011). Optimisation par essaim particulaire de lois de commande
robuste : thorie et mise en uvre pratique, Editions Universitaires Europennes, ISBN
: 978-613-1-59335-2, Saarbrcken, Germany.
[4] Bouallgue, S., Haggge, J., Benrejeb, M., (2010). Structured Loop-Shaping H

Controller Design using Particle Swarm Optimization. Proceedings of the 2010 IEEE
International Conference on Systems, Man, and Cybernetics SMC10, Istanbul.
[5] Bouallgue, S., Haggge, J., Benrejeb, M., (2010). Structured Mixed-Sensitivity H

Design using Particle Swarm Optimization. Proceedings of the 7th IEEE International
Multi-Conference on Systems, Signals and Devices SSD10, Amman.
[6] Bouallgue, S., Haggge, J., Benrejeb, M., (2011). Particle Swarm Optimization-Based
Fixed-Structure H

Control Design. International Journal of Control, Automation, and

Systems 9 (2), pp. 258-266.
[7] Bouallgue, S., Haggge,J., and Benrejeb, M., (2009). Real-Time H

Control Design for

a Thermal Process, Proceedings of the 10th International conference on Sciences and
Techniques of Automatic control & computer engineering STA2009-ACS, pp. 949-958,
Hammamet, Tunisia.
[8] Bhler, H., (1994). Rglage par logique oue, Presses polytechniques et universitaires
romandes, Lausanne.
160 Fuzzy Controllers Recent Advances in Theory and Applications
A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization 23
[9] Eberhart, R.C., Kennedy, J., (1995). A New Optimizer Using Particle Swarm Theory.
Proceedings of the 6th International Symposiumon Micro Machine and Human Science,
Nagoya, pp. 39-43.
[10] Eberhart, R.C., Shi, Y., (2001). Particle Swarm Optimization: Developments,
Applications and Resources. Proceedings of the IEEE Congress on Evolutionary
Computation, Seoul, Korea, pp. 81-86.
[11] Eker, I., Torun, Y., (2006). Fuzzy logic control to be conventional method. Energy
Conversion and Management 47, pp. 377-394.
[12] Eksin, I., Gzelkaya, M., Grleyen, F., (2001). A new methodology for deriving
the rule-base of a fuzzy logic controller with a new internal structure. Engineering
Applications of Articial Intelligence 14, pp. 617-628.
[13] Feedback, (1980). Process Trainer PT326, Feedback Instruments Limited, Crowborough,
United Kingdom.
[14] Goldberg, D.E., (1989). Genetic Algorithms in Search, Optimization, and Machine
Learning. Addison-Wesley.
[15] Grigorie, L. (editor), (2011). Fuzzy Controllers, Theory and Applications, ISBN
978-953-307-543-3, InTech, Croatia.
[16] Gzelkaya, M., Eksin, I., Yesil, E., (2003). Self-tuning of PID-type fuzzy logic controller
coefcients via relative rate observer. Engineering Applications of Articial Intelligence
16, pp. 227-236.
[17] Haggge, J., Ayadi, M., Bouallgue, S., Benrejeb, M., (2010). Design of Fuzzy
Flatness-based Controller for a DC Drive. Control and Intelligent Systems 38 (3), pp.
164-172.
[18] Korosec, P., (editor), (2010). New Achievements in Evolutionary Computation, ISBN
978-953-307-053-7, 2010, InTech, Croatia.
[19] Lazinica, A. (editor), (2009). Particle Swarm Optimization, ISBN 978-953-7619-48-0,
InTech, Croatia.
[20] Lee, C.C., (1990). Fuzzy Logic in Control Systems: Fuzzy Logic Controller. Part I. IEEE
Transactions on Systems, Man, and Cybernetics 20 (2), pp. 404-418.
[21] Lee, C.C., (1990). Fuzzy Logic in Control Systems: Fuzzy Logic Controller. Part II. IEEE
Transactions on Systems, Man, and Cybernetics 20 (2), pp. 419-435.
[22] Mamdani, E.H., (1974). Application of fuzzy logic algorithms for control of simple
dynamic plant, Proceedings of the Institute of Electrical Engineering (IEE), 121 (12),
pp.1585-1588
[23] Mohd, R., Yeoh Keat, H., Sahnius, U., Norhaliza, A.W., (2007). Modelling of PT326
Hot Air Blower Trainer Kit Using PRBS Signal and Cross-Correlation Technique, Jurnal
Teknologi, vol. 42, pp. 9-22.
[24] Passino, K.M., Yurkovich, S., (1998). Fuzzy Control. Addison Wesley Longman, Menlo
Park, California.
[25] Poli, R., Kennedy, J., Blackwell, T., (2007). Particle Swarm Optimization: An Overview.
Swarm Intelligence, Springer 1, pp. 33-57.
[26] Qiao, W.Z., Mizumoto, M., (1996). PID type fuzzy controller and parameters adaptive
method. Fuzzy Sets and Systems 78, pp. 23-35.
[27] Ruben, E.P., Kamran, B., (2007). Particle Swarm Optimization in Structural Design, in:
Chan, F.T.S., Tiwari, M.K. (Eds.), Swarm Intelligence: Focus on Ant and Particle Swarm
Optimization. In-Tech Education and Publishing, Vienna, pp. 373-394.
161 A New Method for Tuning PID-Type Fuzzy Controllers Using Particle Swarm Optimization
24 Will-be-set-by-IN-TECH
[28] Shi, Y., Eberhart, R., (1999). Empirical study of particle swarmoptimization. Proceedings
of the 1999 Congress on Evolutionary Computation, Washington, pp. 1945-1950.
[29] The MathWorks Inc., (2009). Genetic Algorithm and Direct Search ToolboxTM: Users
Guide. Natick.
[30] Van den Bergh, F., (2006). An Analysis of Particle Swarm Optimizers. PhD Thesis,
University of Pretoria, Pretoria, South Africa.
[31] Woo, Z-W., Chung, H-Y., Lin, J-J., (2000). A PID type fuzzy controller with self-tuning
scaling factors. Fuzzy Sets and Systems 115, pp. 321-326.
[32] Yesil, E., Gzelkaya, M., Eksin, I., Tekin, O.A., (2008). Online Tuning of Set-point
Regulator with a Blending Mechanism Using PI Controller, Turk J Elec Engin, 16 (2),
pp. 143-157.
[33] Zadeh, L.A., (1994). Soft computing and fuzzy logic, IEEE Software, 11 (6), pp. 48-56.
[34] Zadeh, L.A., (1965). Fuzzy Sets, Information and Control, 8, pp. 338-353.
162 Fuzzy Controllers Recent Advances in Theory and Applications
1. Introduction
Over the past decades, many advances have been made in the eld of control theory
which rely on state-space theory. The control design methodology that has been most
investigated for the state-feedback control, see for example [1, 2] and the references therein.
The state-feedback control designsupposes that all the systemstates are available, which is not
always possible in realistic applications. Instead, one has to deal with the absence of full-state
information by using observers. From the control point of view, observers can be used as
part of dynamical controllers. This observer-based design has been extensively studied in
the literature [3, 4]. However, it leads to high-order controllers. As a matter of fact, one has
to solve a large problem, which increases numerical computations for large scale systems.
Other difculties may arise, if we consider additional performances, such as disturbance
rejection, time delays, uncertainties, etc. Hence, it is more suitable to develop methodologies
which involve a design with a low dimensionality. In this context, intensive efforts have
been devoted to design low-order controllers [3, 57]. In particular, it has been shown
that designing reduced order stabilizing controllers can be cast as a static output-feedback
stabilization problem. Also, it is recognized that, in general, the static output-feedback
control design may not exist for certain systems. Note that an important advantage of these
controllers is that they are easy to implement without signicant numerical burden.
In general, the synthesis of static output-feedback stabilizing controllers is known to be
a hard task [57]. The main difculty rises from its nonconvexity. In the literature,
some convexication techniques and iterative algorithms have been proposed to handle this
problem [3, 5, 7]. A comprehensive survey on static output-feedback stabilization can be
found in [6]. The authors show that despite the considerable efforts devoted to solve this
problem, there is yet no methodology that can solve it exactly, so it is still an important open
topic. However, it has been shown that for SISO (Single-Input Single-Output) systems, this
problem can be solved exactly based on an algebraic characterization [8, 9]. Unfortunately,
these approaches are valid only for SISO case and cannot be used to take into account
additional constraints on the system. In any case, the investigation of this topic within the
eld of fuzzy control is continuously increasing and leading to many approaches. A most
Output Tracking Control for Fuzzy Systems
via Static-Output Feedback Design
Meriem Nachidi and Ahmed El Hajjaji
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/47138

Chapter 7
2 Will-be-set-by-IN-TECH
efcient approach is based on the Linear Matrix Inequality (LMI) technique: see for example
[1012]. Indeed, since the developed interior-point methods [13], LMIs can be solved in
polynomial-time, using numerical algorithms [14]. Recently, other approach, based on a
projective algorithm has been proposed [15]. Notice that the existing LMI tools have opened
an important research area in system and control theory and tackled numerous unsolved
problems [14]. Therefore, our main focus in this chapter is the design of static output-feedback
controllers using LMI theory for a class of nonlinear systems describedby Takagi-Sugeno (T-S)
fuzzy models.
Recently, the study of T-S fuzzy models has attracted the attention of many of researchers: see
[16] and references therein. Fuzzy models have local dynamics (i.e., dynamics in different
state space regions), that are represented by local linear systems. The overall model of
a fuzzy system is then obtained by interpolating these linear models through nonlinear
fuzzy membership functions. Unlike conventional modeling techniques, which use a single
model to describe the global behavior of a nonlinear system, fuzzy modeling is essentially a
multi-model approach, in which simple local linear submodels are designed in the form of a
convex combination of local models in order to describe the global behavior of the nonlinear
system. This kind of models has proved to be a good representation for a certain class of
nonlinear dynamic systems.
Since the work by [17] on stability analysis and state feedback stabilization for fuzzy systems,
the Parallel Distributed Compensation (PDC) procedure has extensively been used for the
control of such systems: for more details see [16]. The basic idea of this procedure is to design
a feedback gain for each local model, and then to construct a global controller from these
local gains, so that the global stability of the overall fuzzy system can be guaranteed. The
most interesting of this concept is that the obtained stability conditions do not depend on
the nonlinearities (membership functions), so that this makes possible to use linear system
techniques for nonlinear control design.
Up to now, the stabilization control design for T-S systems is successfully investigated based
on state-feedback or static/dynamic output-feedback [18, 19]. However, the design of a
controller which guarantees an adequate tracking performance for nite-dimensional systems
is more general problem than the stabilization one, and is still attract considerable attentions
due to demand from practical dynamical processes in electric, mechanics, agriculture, . . . .
One of our main interest in this chapter is solving the static output-feedback tracking
problem. Due to the fact that the T-S fuzzy models aggregate a set of local linear subsystems,
blended together through nonlinear scalar functions, the static output-feedback control
problem can be very complicated to solve. With regard to the literature of fuzzy control,
a few recent approaches have dealt with the tracking control design problem for nonlinear
systems described by T-S fuzzy model. Generally speaking, the incorporation of linearization
techniques and adaptive schemes usually needs systems perfect knowledge and leads to
complicated adaptation control laws. In [20], the author has been shown that the use of the
feedback linearization strategy [21] may lead to unbounded controllers, since their stability
is not guaranteed. To overcome these drawbacks, LMI-based methodologies have been
developed for tracking control problem, using observer-based fuzzy controller to deal with
the absence of full-state information [22].
In this context, this chapter will tackle the static output-feedback fuzzy tracking control
problem, focusing on an H

tracking performance, related to an output tracking error for

all bounded references inputs. The presented results are an extension of already published
works for the stabilization case [12, 23]. In fact, to solve the nonconvexity problem, inherent
164 Fuzzy Controllers Recent Advances in Theory and Applications
Output Tracking Control for Fuzzy Systems Via Static-Output Feedback Design 3
to static output-feedback control synthesis, a cone complementarity formulation [7] for T-S
fuzzy systems is used combined with an iterative algorithm. This algorithm has to optimize a
linear objective function subject to a set of LMIs in each iteration. Thus, controllers are derived
that not only ensure stability of the closed-loop system, but also provide a prescribed level of
output tracking error attenuation.
The main contribution of this chapter is the purpose of a simple procedure reected by
an efcient LMI-based iterative algorithm to solve the fuzzy tracking control problem for
nonlinear systems described by discrete-time T-S models. Therefore, since the proposed fuzzy
tracking controllers have a low-order character, they are suitable for industrial application.
Furthermore, this chapter shows an application to a relevant practical problem, in power
engineering and drives eld, of the proposed design procedure: guaranteeing a good tracking
of the output voltage of DC-DC buck converter [2426].
2. Problem formulation and preliminaries
Consider a nonlinear system which is approximated by a T-S fuzzy model of the following
form:
i
th
Rule: IF z
1
(k) is
i
1
and . . . and z
p
(k) is
i
p
,
THEN

x(k +1) = (A
i
+A
i
(k))x(k) + (B
i
+B
i
(k))u(k) + E
i
w(k),
y(k) = C
i
x(k), i = 1, . . . , N,
(1)
where x(k)
n
is the state vector, u(k)
n
u
is the input vector, w(k)
n
w
comprises
the bounded external disturbances and y(k)
n
y
is the system output. N is the number of
IF-THEN rules. z
1
(k), . . . z
p
(k) are the premise variables (that comprises states and/or inputs)
and
i
j
(i = 1, . . . , N, j = 1, . . . , p) are the fuzzy sets. A
i
, B
i
, C
i
and E
i
are known constant
matrices of appropriate size, A
i
(k), B
i
(k) are unknown matrices representing time-varying
parameter uncertainties, and are assumed to be as follows:
[A
i
(k) B
i
(k)] = [M
1
F(k)N
1i
M
2
F(k)N
2i
], i = 1, 2, . . . , N, (2)
where M
i
, N
1i
and N
2i
are known real constant matrices. F(k) is the uncertainty function that
satises the classical bounded condition:
F(k)
T
F(k) I, k. (3)
Thus, the global T-S model is an interpolation of all uncertain subsystems through nonlinear
functions [16]:
x(k +1) =
N

i=1

i
(z) [(A
i
+A
i
(k))x(k) + (B
i
+B
i
(k))u(k) + E
i
w(k)]
N

i=1

i
(z)
,
=
N

i=1

i
(z) [(A
i
+A
i
(k))x(k) + (B
i
+B
i
(k))u(k) + E
i
w(k)] ,
y(k) =
N

i=1

i
(z)C
i
x(k),
(4)
165 Output Tracking Control for Fuzzy Systems via Static-Output Feedback Design
4 Will-be-set-by-IN-TECH
where
i
, i = 1, . . . , N, is the membership function corresponding to systemrule i, and
i
(z) =

i
(z)/
N

i=1

i
(z), fullls the convex property: 0
i
(z) 1 and
N

i=1

i
(z) = 1.
Note that using the so-called sector of nonlinearity approach, a wide number of nonlinear
systems can be representedexactly by T-S models in a compact set of the state space. However,
with the growing complexity of nonlinear systems, it is useful to take into account the
approximations in the dynamical process. Thus, the main objective of the next paragraph
is to provide stability conditions that ensure the tracking performance for the uncertain T-S
model (4).
3. H

output tracking performance analysis

This section gives sufcient stability conditions which ensure an H

output tracking
performance of the uncertain system (4) using a fuzzy Lyapunov function. We recall the
following lemma which will be used in this section.
Lemma 3.1. [27] Let A, D, S, W and F be real matrices of appropriate dimension such that
W > 0 and FF
T
I. Then, for any scalar > 0 such that W DD
T
> 0, we have
(A+DFS)
T
W
1
(A+DFS) A
T
(W DD
T
)
1
A+
1
S
T
S.
Suppose that the desired trajectory can be generated by the following reference model as
follows:

x
d
(k +1) = Ax
d
(k) +Br(k),
y
d
(k) = Cx
d
(k),
(5)
where, y
d
(k) has the same dimension as y(k), x
d
(k) and r(k)
n
r
are respectively the
reference state and the bounded reference input, A, B and C are appropriately dimensional
constant matrices with AHurwitz.
Since we deal with the static output-feedback control design problem, the fuzzy controller
can incorporates information from y(k) and y
d
(k). Thus, the control law which is based on
the classical structure of the Parallel Distributed Compensation (PDC) concept [17, 28] shares
the same fuzzy sets as the T-S system and can be given as follows:
i
th
Rule: IF z
1
(k) is
i
1
and . . . and z
p
(k) is
i
p
,
THEN u(k) = K
i
(y(k) y
d
(k)),
(6)
where the the controller gain K
i
is to be chosen. The overall static output-feedback control law
is thus inferred as:
u(k) =
N

i=1

i
(z)K
i
(y(k) y
d
(k)). (7)
The advantages of the static output-feedback controller (7), is well discussed in the literature
[3], [6]. This fact motivates us to use such type of control law avoiding the complex control
schemes with an additional observer.
166 Fuzzy Controllers Recent Advances in Theory and Applications
Output Tracking Control for Fuzzy Systems Via Static-Output Feedback Design 5
Combining (4), (5) and (7), the following augmented closed-loop system is obtained
x(k +1) =
N

i,j,s=1

i
(z)
j
(z)
s
(z)
_
(G
1ijs
+ G
2ijs
(k)) x(k) + W
i
w(k)
_
,
(8)
where
G
1ijs
=

A
i
+ B
i
K
j
C
s
B
i
K
j
C
0 A

,
G
2ijs
(k) =

A
i
(k) +B
i
(k)K
j
C
s
B
i
(k)K
j
C
0 0

,
W
i
=

E
i
0
0 B

, x =

x(k)
x
d
(k)

, w =

w(k)
r(k)

,
(9)
Hence, to meet the required tracking performance, the effect of w(k) on the tracking error
y(k) y
d
(k) should be attenuated below a desired level in the sense of [29]:
k
f

k=0
(y(k) y
d
(k))
T
(y(k) y
d
(k))
2
k
f

k=0
w(k)
T
w(k),
(10)
k
f
= 0, and w(k) l
2
, k
f
is the control nal time.
The following theorem shows that H

output tracking performances can be guaranteed if

there exist some matrices satisfying certain conditions.
Theorem 3.1. The augmented closed-loop system in (8) achieves the H

output tracking performance

, if there exists matrices P
1
> 0, . . . , P
N
> 0 and controller gains K
1
, . . . , K
N
such that the
following conditions hold:

P
1
r
0 0 G
1ijs
W
i

M
0 I 0

N
ijs
0 0
0 0 I H
i
0 0
G
T
1ijs

N
T
ijs
H
T
i
P
i
0 0
W
T
i
0 0 0
2
I 0

M
T
0 0 0 0
1
I

< 0, 1 i, j, s, r N, (11)
167 Output Tracking Control for Fuzzy Systems via Static-Output Feedback Design
6 Will-be-set-by-IN-TECH
where
G
1ijs
and W
i
are dened in (9), H
i
= [C
i
C] ,

M =

M
1
M
2
0 0

and

N
ijs
=

N
1i
0
N
2i
K
j
C
s
N
2i
K
j
C

.
Proof. Consider the following fuzzy Lyapunov function V( x, k) given by
V( x, k) = x(k)
T
N

i=1

i
(z)P
i
x(k).
The stability of (8) is ensured, under zero initial condition, with guaranteed H

performance
(10) if [29]:
V( x, k) + (y(k) y
d
(k))
T
(y(k) y
d
(k))
2
w(k)
T
w(k) < 0 (12)
where V( x, k) is the rate of V along the trajectory:
V( x, k) = V( x(k +1)) V( x(k)). (13)
By substituting (13) in(12), we have:
x(k +1)
T
P
+
x(k +1) x(k)
T
P
z
x(k) + (y(k) y
d
(k))
T
(y(k) y
d
(k))
2
w(k)
T
w(k) < 0, (14)
where
P
z
=
N

i=1

i
(z)P
i
and P
+
=
N

i=1

i
(z(k +1))P
i
.
Now, let
G
z
(k) =
N

i,j,s=1

i
(z)
j
(z)
s
(z)G
1ijs
+
N

i,j,s=1

i
(z)
j
(z)
s
(z)G
2ijs
(k),
W
z
=
N

i=1

i
(z)W
i
.
(15)
Then, the inequality (14) can be rewritten as follows
[G
z
(k) x(k) + W
z
w(k)]
T
P
+
[G
z
(k) x(k) + W
z
w(k)] x(k)
T
P
z
x(k)
2
w(k)
T
w(k)+
(y(k) y
d
(k))
T
(y(k) y
d
(k)) < 0.
(16)
By consequence, (16) leads to:

x(k)
w(k)

T
(M
1
M
2
)

x(k)
w(k)

< 0,
(17)
168 Fuzzy Controllers Recent Advances in Theory and Applications
Output Tracking Control for Fuzzy Systems Via Static-Output Feedback Design 7
where
M
1
=

G
z
(k)
T
P
+
G
z
(k) G
z
(k)
T
P
+
W
z
W
T
z
P
+
G
z
(k) W
T
z
P
+
W
z

,
M
2
=

P
z
H
T
z
H
z
0
0
2

,
H
z
=
N

i=1

i
(z)H
i
.
(18)
Thus, to proof (12), it is sufcient to show that
M
1
M
2
< 0. (19)
The rst part of (19) can also be rewritten as
M
1
M
2
= (

G
z
+

MF(k)N
z
)
T
P
+
(

G
z
+

MF(k)N
z
), (20)
where

G
z
=
_
G
1z
W
z

, G
1z
=
N

i,j,s=1

i
(z)
j
(z)
s
(z)G
1ijs
,
and N
z
=
_
N

i,j,s=1

i
(z)
j
(z)
s
(z)

N
1ijs
0
_
.
(21)
On the other hand, pre- and post-multiplying (11) by diag{P
r
, I, I, I, I, I} gives

r
ijs

P
r
0 0 P
r
G
1ijs
P
r
W
i
P
r

M
0 I 0

N
ijs
0 0
0 0 I H
i
0 0
G
T
1ijs
P
r

N
T
ijs
H
T
i
P
i
0 0
W
T
i
P
r
0 0 0
2
I 0

M
T
P
r
0 0 0 0
1
I

< 0, 1 i, j, s, r N. (22)
169 Output Tracking Control for Fuzzy Systems via Static-Output Feedback Design
8 Will-be-set-by-IN-TECH
Since
N

i=1

i
(z) =
N

r=1

r
(k +1) = 1, (22) can be written as
N

r=1

r
(k +1)
N

i,j,s=1

i
(z)
j
(z)
s
(z)
r
ijs

P
+
0 0 P
+
G
1z
P
+
W
z
P
+

M
0 I 0

N
z
0 0
0 0 I H
z
0 0
G
T
1z
P
+

N
T
z
H
T
z
P
z
0 0
W
T
z
P
+
0 0 0
2
I 0

M
T
P
+
0 0 0 0
1
I

< 0, (23)
Applying Schur complement on (23), it is straightforward to verify that the condition (23) is
equivalent to the following inequalities:
(P
+

G
z
)
T
_
P
+
P
+

M

M
T
P
+
_
1
P
+

G
z
+
1
N
T
z
N
z
M
2
< 0 and
P
+
P
+

M

M
T
P
+
> 0.
(24)
Using (20), (24) and Lemma 3.1, we have
M
1
M
2
= (

G
z
+

MF(k)N
z
)
T
P
+
(

G
z
+

MF(k)N
z
)
(P
+

G
z
)
T
_
P
+
P
+

M

M
T
P
+
_
1
P
+

G
z
+
1
N
T
z
N
z
M
2
< 0.
(25)
By consequence
k
f

k=0
(y(k) y
d
(k))
T
(y(k) y
d
(k)) <
2
k
f

k=0
w(k)
T
w(k).
Hence, H

output tracking performance is achieved with the prescribed attenuation level .

On the other hand, it follows from(11) and (25) that V( x) < 0 for w(k) = 0, which leads that
the uncertain system (8) with w(k) = 0 is robustly asymptotically stable.
4. H

fuzzy tracking controller synthesis

In this section, a cone complementarity formulation [7] is used to solve the bilinearity involved
in (11). The idea is based on converting the conditions (11) to convex and nonconvex parts and
then casting them into an optimization problem subject to some LMIs. For this, rst recall the
following lemma, which generalizes the result of [7].
Lemma 4.1. [12] Let P
i

nn
, Q
i

nn
, i = 1, . . . , N be any symmetric positive denite
matrices, then the following statements are equivalent:
170 Fuzzy Controllers Recent Advances in Theory and Applications
Output Tracking Control for Fuzzy Systems Via Static-Output Feedback Design 9
(a): P
i
Q
i
= I, i = 1, . . . , N.
(b):

i=1
Tr(P
i
Q
i
) = N n,

P
i
I
I Q
i

0, 1 i N.
Using P
r
= Q
1
r
, the stability condition (11) can be rewritten as follows:

r
ijs

Q
r
0 0 G
1ijs
W
i

M
0 I 0

N
ijs
0 0
0 0 I H
i
0 0
G
T
1ijs

N
T
ijs
H
T
i
P
i
0 0
W
T
i
0 0 0
2
I 0

M
T
0 0 0 0
1
I

< 0, 1 i, j, s, r N, (26)
P
r
Q
r
= I, 1 r N. (27)
Before giving the nal formulation of the problem in hand, we suggest to relax the LMIs (26)
fromthe point of viewnumber of LMIs to be satised, for this, we suggest to use the following
lemma.
Lemma 4.2. [12] Consider the following matrix

A =
N

i,j,s=1

ijs
A
ijs
, where
ijs
=
i

s
and
N

i=1

i
=
1. Then,

A can be expressed as follows

A =
N

i=1

3
i
A
iii
+
N

s>ji

ijs
(A
ijs
+ A
jsi
+ A
sij
) +
N

sj>i

ijs
(A
sji
+ A
isj
+ A
jis
),
Moreover,
N

ijs=1

ijs
=
N

i=1

3
i
+3
N

s>ji

ijs
+3
N

sj>i

ijs
= 1.
Hence, using Lemma 4.2, (26) can be rewritten as follows:

r
iii
< 0, 1 i, r N,

r
ijs
0, 1 i j < s N, 1 r N,

r
ijs
0, 1 i < j s N, 1 r N,
(28)
171 Output Tracking Control for Fuzzy Systems via Static-Output Feedback Design
10 Will-be-set-by-IN-TECH
where,

r
iii

Q
r
0 0 G
1iii
W
i

M
0 I 0

N
iii
0 0
0 0 I H
i
0 0
G
T
1iii

N
T
iii
H
T
i
P
i
0 0
W
T
i
0 0 0
2
I 0

M
T
0 0 0 0
1
I

r
ijs

3Q
r
0 0 G
ijs
+ G
jsi
+ G
sij
W 3

M
0 3I 0

N
ijs
+

N
jsi
+

N
sij
0 0
0 0 3I H
i
+ H
j
+ H
s
0 0
(P
i
+ P
j
+ P
s
) 0 0
0 0 0 3
2
I 0
0 0 0 0 3
1
I

r
ijs

3Q
r
0 0 G
sji
+ G
isj
+ G
jis
W 3

M
0 3I 0

N
sji
+

N
isj
+

N
jis
0 0
0 0 3I H
i
+ H
j
+ H
s
0 0
(P
i
+ P
j
+ P
s
) 0 0
0 0 0 3
2
I 0
0 0 0 0 3
1
I

,
where W = W
i
+ W
j
+ W
s
.
From Lemma 4.2, It is only sufcient to see that [12]
N

i=1

i
(z)
r
ijs
=
N

i=1

3
i
(k)
r
iii
+
N

ij<s

i
(z)
j
(z)
s
(z)
r
ijs
+
N

i<js

i
(z)
j
(z)
s
(z)
r
ijs
.
It should be noted that, Lemma 4.2 is very useful in reducing the number of LMIs to be
satised. Indeed, (26) leads to N
4
LMIs to be satised. In contrast, by using Lemma 4.2,
this number decreases to (N
2
(N
2
+2))/3.
172 Fuzzy Controllers Recent Advances in Theory and Applications
Output Tracking Control for Fuzzy Systems Via Static-Output Feedback Design 11
Now, back to our main problem. We suggest to use Lemma 4.1 to handle the nonconvexity
involved in (27), as it is clearly shown by the following theorem:
Theorem 4.1. Given a weight > 0 and > 0. The augmented closed-loop system in (8) achieves
the H

output tracking performance , if there exists positive denite matrices P

1
> 0, . . . , P
N
> 0,
Q
1
> 0, . . . , Q
N
> 0 and controller gains K
1
, . . . , K
N
such that the following optimization problem is
solvable and equal to n
x
N:

minimize
K
i
,P
i
,Q
i
,

i=1
Tr(P
i
Q
i
) + (1 )
subject to:
(28) and

P
i
I
I Q
i

0, 1 i N.
(29)
The following iterative algorithm [7, 12] can be used to linearize the objective function of the
optimization problem (29).
Algorithm 4.1
give a weight , x a tolerance (for example = 10
6
) and execute the following steps:
Step 1: Set P
0
i
= I and Q
0
i
= I, for i = 1, . . . , N.
Step 2: Solve the following LMI optimization:
minimize
K
i
,P
i
,Q
i
,

i=1
Tr(P

i
Q
i
+ Q

i
P
i
) + (1 )
subject to :
(28) and

P
i
I
I Q
i

0, 1 i N.
Step 3: If P
i
Q
1
i
.
While P
i
Q
1
i
,
Select = 0.01 and repeat from step 1. Else
Set P

i
P
i
, Q

i
Q
i
and repeat from step 2.
Remark 4.1. In the optimization problem (29), the attenuation level is also included in the
optimization function. Thus, a multi-objective optimization problem is solved by the Algorithm 4.1.
5. Illustrative example
In this section, the proposed tracking control scheme is applied to regulate the output voltage
of DC-DC converter. The model of a buck converter is described in Fig. 1. Using the Kirchoff
laws, the converter of Fig. 1 can be representedby the following discrete-time nonlinear model
[24]:
173 Output Tracking Control for Fuzzy Systems via Static-Output Feedback Design
12 Will-be-set-by-IN-TECH
Figure 1. Buck converter circuit.
x(k +1) =

T
s
L
(R
L
+
R(k)R
c
R(k)+R
c
) +1
T
s
R(k)
L(R(k)+R
c
)
T
s
R(k)
C(R(k)+R
c
)
T
s
C(R(k)+R
c
)
+1

x(k)+

T
s
L
(R
M
i
L
(k) V
in
(k) V
D
)
0

u(k) +

T
s
V
D
L
0

,
y(k) =
_
R(k)R
c
(R(k)+R
c
)
R(k)
(R(k)+R
c
)
_
x(k),
(30)
where x(k) = [i
L
(k) v
c
(k)]
T
is the state vector, u(k) is the control vector i.e. the duty cycle
of the switched M, y(k) is the output vector i.e. the output voltage and T
s
is the sampling
period T
s
= 0.001 1/f
0
, with f
0
is the resonance frequency of the buck converter (30). R(k)
and V
in
(k) are uncertain parameters satisfying R(k) [R(k), R(k)], V
in
(k) [V
in
(k), V
in
(k)].
Table (1) gives the parameter values of the buck converter (Fig. 1). Similar to [24], we assume
that the inductor current belongs in a compact set: i
L
(k) [i
L
, i
L
], and select the membership
functions as follows

1
(k) =
i
L
(k) + i
L
i
L
i
L
,
2
(k) = 1
1
(k). (31)
The nonlinear system (30) can be represented by the following uncertain T-S model:
Rule
i
If i
L
(k) is
i
Then

x(k +1) = (A
noi
+A
i
(k))x(k) + (B
noi
+B
i
(k))u(k) + E
i
w(k),
y(k) = C
i
x(k), i = 1, 2,
(32)
174 Fuzzy Controllers Recent Advances in Theory and Applications
Output Tracking Control for Fuzzy Systems Via Static-Output Feedback Design 13
where
A
no1
= A
no2
=
A
1
+A
1
2
, B
no1
=
B
1
+B
1
2
, B
no2
=
B
2
+B
2
2
, with
A
1
= A
2
=

T
s
L
(R
L
+
RR
c
(R+R
c
)
) +1
T
s
R
L(R+R
C
)
T
s
R
C(R+R
C
)

T
s
C(R+R
c
)
+1

,
A
1
= A
2
=

T
s
L
(R
L
+
RR
c
(R+R
c
)
) +1
T
s
R
L(R+R
C
)
T
s
R
C(R+R
C
)

T
s
C(R+R
c
)
+1

,
B
1
=

T
s
L
(R
M
i
L
V
in
V
D
)
0

, B
2
=

T
s
L
(R
M
i
L
V
in
V
D
)
0

,
B
1
=

T
s
L
(R
M
i
L
V
in
V
D
)
0

, B
2
=

T
s
L
(R
M
i
L
V
in
V
D
)
0

,
C
1
= C
2
=
_
RR
c
R+R
c
R
R+R
c
_
, and E
1
= E
2
=
_
1
0
_
.
A
1
(k), A
2
(k), B
1
(k) and B
2
(k) can be represented in the form of (2) with M
1
=
0.1, M
2
=

1 0
0 0

, N
11
= 10
A
1
A
1
2
, N
12
= N
11
, N
21
=
B
1
B
1
2
, N
22
=
B
2
B
2
2
.
In this example, the objective is to make the output voltage of the buck converter, i.e. v
o
follow
a desired signal to meet the H

tracking performance of the uncertain system (30).

The reference system matrices of (5) is selected as follows
A =

0.5 0
0 0.5

, B =

0
1

, C =
_
0 1

. (33)
Let = 0.99 and = 1, using the Algorithm 4.1, the following feasible solution is obtained
after only 41 iterations:
P
1
=

0.122328 0.72818 0 0.070451

0.72818 7.378511 0 0.550637
0 0 1 0
0.070451 0.550637 0 2.846761

,
175 Output Tracking Control for Fuzzy Systems via Static-Output Feedback Design
14 Will-be-set-by-IN-TECH
P
2
=

0.124832 0.7441565 0 0.09171

0.7441565 7.455168 0 0.661751
0 0 1.116002 0
0.09171 0.661751 0 3.00123

,
Q
1
=

19.852249 1.950697 0 0.113988

1.950697 0.329190 0 0.015398
0 0 1 0
0.113988 0.015398 0 0.357075

,
Q
2
=

19.869471 1.967944 0 0.173243

1.967944 0.3317250 0 0.013007
0 0 0.896056 0
0.173243 0.013007 0 0.341358

,
K
1
= 6.0943; K
2
= 7.1963,
and the H

output tracking performance index: = 2.52. Hence, according to (7), the static
output-feedback control law that ensures the desired trajectory tracking for (30) is given as
follows:
u(k) = (
1
(k)K
1
+
2
(k)K
2
)(y(k) y
d
(k)). (34)
Fig. 2 shows the evolution of the output signal of the nonlinear system (30), using the fuzzy
controller, with an external disturbance input w(k) dened as w(k) =
r
o
1+15(k+1)
T
s
V
D
/L,
where, r
o
is a random number taken from a uniform distribution over [0, 2], the uncertain
parameters are as follow
R(k) =
R+R
2
+
RR
2
cos(k/T
s
),
V
in
(k) =
V
in
+V
in
2
+
V
in
V
in
2
cos(k/T
s
),
(35)
and the reference signal r(k), are supposed to be

r(k) = 12V for 0 k 0.005s,

r(k) = 6V for 0.005 < k 0.01s,
r(k) = 24V for k > 0.01s,
(36)
176 Fuzzy Controllers Recent Advances in Theory and Applications
Output Tracking Control for Fuzzy Systems Via Static-Output Feedback Design 15
0 0.005 0.01 0.015
0
5
10
15
20
25
time (s)
y
d
(k)
y(k)
Figure 2. Response of y(k) and y
d
(k).
Fig. 3 and Fig. 4 depict a zoom of Fig. 2 at 0 s and between 5 ms and 10 ms respectively. It can
be seen that the designed fuzzy static output-feedback controller ensures the robust stability
of the nonlinear system(30) and guarantees an acceptable H

trajectory tracking performance

level.
0 1 2
x 10
4
0
2
4
6
8
10
12
14
16
18
time (s)
y
d
(k)
y(k)
Figure 3. Zoom on Fig. 2 at 0 sec.
177 Output Tracking Control for Fuzzy Systems via Static-Output Feedback Design
16 Will-be-set-by-IN-TECH
5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10
x 10
3
5.8
5.85
5.9
5.95
6
6.05
6.1
6.15
6.2
time (s)
y
d
(k)
y(k)
Figure 4. Zoom on Fig. 2 between 5 msec and 10 msec.
Parameter Value Unity
Input voltage, V
in
(k) V
in
(k) [10, 30] V
Current in the inductance , i
L
-8 - 8 A
Inductance, L 98.58 H
Parasitic resistance of L, R
L
48.5 m
Capacitor, C 202.5 F
Parasitic resistance of C, R
c
162 m
Resistance of Switch, R
M
0.27
Diode voltage, V
D
0.82 V
Load resistance, R(k) R(k) [2, 10]
Table 1. Parameter values of the buck converter.
6. Conclusion
In this chapter, the problem of model reference tracking control with a guaranteed H

performance is solved for uncertain discrete-time fuzzy systems. Based on the fuzzy
Lyapunov function and cone complementary formulation, a fuzzy static output controller is
calculated to make small as possible as the tracking output error and reject disturbances.
178 Fuzzy Controllers Recent Advances in Theory and Applications
Output Tracking Control for Fuzzy Systems Via Static-Output Feedback Design 17
Author details
Meriem Nachidi
Departamento de Ingeneria de Sistemas y Automtica, University of Carlos III, Madrid, Spain.
Ahmed El Hajjaji
Laboratoire Modlisation, Information et Systmes, University of Picardie Jules Verne, Amiens, France.
7. References
[1] B.D.O. Anderson and J.B. Moore. Linear Optimal Control. Prentice-Hall, Englewood Cliffs,
1971.
[2] H. Khalil. Nonlinear Systems. Pearson Higher Education, 2002.
[3] J. C. Geromel, C. C. de Souza, and R. E. Skelton. Static output feedback controllers:
Stability and convexity. IEEE Transactions on Automatic Control, 43, 1998.
[4] R. Johansson and A. Robertsson. Observer-basedstrict positive real (spr)feedback control
system design. Automatica, 38, 2002.
[5] T. Iwasaki and R.E. Skelton. The xy-centring algorithm for the dual lmi problem: a new
approach to xed-order control design. Int. J. Control, 62(6):12571272, 1995.
[6] V.L. Syrmos, C.T. Abdallah, P. Dorato, and K. Grigoriadis. Static output feedback

Ua
survey. Automatica, 33(2):125137, 1997.
[7] L. El Ghaoui, F. Oustry, and M. Ait Rami. A cone complementarity linearisation
algorithm for static output-feedback and related problems. IEEE Trans. Automat. Contr.,
42:1171, 1997.
[8] D. Mustafa. LQG optimal scalar static output feedback. Systems and Control Letters,
27(1):919, 1996.
[9] D. Henrion, M. Sebek, and V. Kucera. An algorithm for static output feedback
simultaneous stabilization of scalar plants. In Proceedings of the IFAC World Congress on
Automatic Control, Barcelona, Spain, 2001.
[10] S.-S. Chen, Y.-C. Chang, S.-F. Su, S.-L. Chung, and T.-T. Lee. Robust static
output-feedback stabilization for nonlinear discrete-time systems with time delay via
fuzzy control approach. IEEE Transactions on Fuzzy Systems, 13(2):263272, 2005.
[11] D. Huang and S. K. Nguang. Robust H

static output feedback control of fuzzy

systems: an ilmi approach. IEEE Transactions on Systems, Man, and Cybernetics, Part-B,
36(1):216222, 2006.
[12] M. Nachidi, A. Benzaouia, F. Tadeo, and M. Ait Rami. LMI-based approach for
output-feedback stabilization for discrete time Takagi-Sugeno systems. IEEE Transactions
on Fuzzy Systems, 16(5):11881196, 2008.
[13] Yu. Nesterov and A. Nemirovsky. Interior-Point Polynomial Methods in Convex
Programming. Philadelphia, PA: SIAM, 1994.
[14] S. Boyd, L. El Ghaoui, E. Feron, and V. Balakrishnan. Linear matrix inequalities in system
and control theory. Philadelphia, PA: SIAM, 1994.
[15] M. Ait Rami, U. Helmeke, and J. B. Moore. A nite steps algorithm for solving convex
feasability problems. Journal of Global Optimization, 38:143160, 2007.
[16] K. Tanaka and H. O. Wang. Fuzzy Control Systems Design and Analysis: A Linear Matrix
Inequality Approach. New York: Wiley , New York: Wiley, 2001.
[17] K. Tanaka and M. Sugeno. Stability analysis and design of fuzzy control systems. Fuzzy
Sets and Syst., 45:135156, 1992.
179 Output Tracking Control for Fuzzy Systems via Static-Output Feedback Design
18 Will-be-set-by-IN-TECH
[18] T. M. Guerra and L. Vermeiren. LMI-based relaxed nonquadratic stabilization conditions
for nonlinear systems in the takagi-sugenos form. Automatica, 40(5):823829, 2004.
[19] S. Xu and J. Lam. Robust H

control for uncertain discrete-time-delay fuzzy systems via

output feedback controllers. IEEE Trans. Fuzzy Syst., 13:8293, 2005.
[20] H. Ying. Analytical analysis and feedback linearization tracking control of the general
takagi-sugeno fuzzy dynamic systems. IEEE Transactions Systems, Man, Cybern.,
29:290298, 1999.
[21] C. C. Kung and H. Li. Tracking control of nonlinear systems by fuzzy model-based
controller. In 6th IEEE International Conference, page 623628, 1997.
[22] C. S. Tseng, B. S. Chen, and H. J. Uang. Fuzzy tracking control design for nonlinear
dynamic systems via T-S fuzzy model. IEEE Transaction on fuzzy systems, 8(2):200211,
2001.
[23] M. Nachidi, F. Tadeo, A. Benzaouia and M. Ait Rami. Static output-feedback for
Takagi-Sugeno systems with delays. International Journal of Adaptive Control and Signal
Processing, 25:295312, 2011.
[24] K.-Y. Lian, J.-J. Liou, and C.-Y Huang. LMI-based integral fuzzy control of DC-DC
converters. IEEE Transactions on Fuzzy systems, 14:7180, 2006.
[25] A. Balestrino, A. Landi, and L. Sani. Cuk converter global control via fuzzy logic and
scaling factors. IEEE Trans. Indusriel Applications, 38:406413, 2002.
[26] G. Papafotiou, T. Geyer, and M. Morari. Hybrid modelling and optimal control of
swich-mode dc-dc converters. In IEEE Workshop on Computers in Power Electronics,
Champaign, IL, USA, 2004.
[27] Y. Wang, L. Xie, and C. E. de Souza. Robust control of a class of uncertain nonlinear
systems. Syst. Control Lett., 19, 1999.
[28] H. O. Wang, K. Tanaka, and M. F. Grifn. An approach to fuzzy control of nonlinear
systems: Stability and design issues. IEEE Trans. Fuzzy syst., 4(1):1423, 1996.
[29] B.-S. Chen, C.-S. Tseng, and H.-J. Uang. Robustness design of nonlinear dynamic systems
via fuzzy linear control. IEEE Trans. Fuzzy Syst., 7:571585, 1999.
180 Fuzzy Controllers Recent Advances in Theory and Applications
Chapter 8
FPGA-Based Motion Control IC for Linear Motor
Drive X-Y Table Using Adaptive Fuzzy Control
Ying-Shieh Kung, Chung-Chun Huang and Liang-Chiao Huang
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/48457
1. Introduction
The development of a compact and high performance motion controller for the X-Y table of
a CNC machine has been an important field in literatures (Groove, 1996; Goto et al., 1996;
Hanafi et al., 2003). The typical architecture of the conventional motion control system for X-
Y table is shown in Fig. 1, which consists of a central controller, two sets of servo drivers and
an X-Y table. The central controller, which usually adopts a float-pointed processor,
performs the function of motion trajectory and data communication with servo drivers and
with external device. Each servo driver usually use a fixed-pointed processor, some specific
ICs and an inverter to perform the functions of position/speed/current control at each single
axis of X-Y table and to do the data communication with the central controller. Data
communication between two devices uses an analog signal, a bus signal or a serial
asynchronous signal. However, the motion control system in Fig.1 has some drawbacks,
such as large volume, easy effect by the noise, expensive cost, inflexible, etc. In addition,
data communication and handshake protocol between the central controller and servo
drivers slow down the system executing speed.
In recent years, the FPGA has been widely applied in implementing the digital control
system (Cho, 2009; Monmasson et al. 2011; Sanchez-Solano et al. 2007). Besides, an
embedded processor IP and an application IP can be developed and downloaded into FPGA
to construct a SoPC environment (Altera, 2004; Hall and Hamblem, 2004), allowing the users
to design a SoPC module by mixing hardware and software in one FPGA chip (Kung et al.
2004; Kung and Tsai, 2007; Kung and Chen, 2008). Therefore, based on the FPGA
technology, we improve the aforementioned drawbacks and integrate the central controller
and the controller part of two servo drivers in Fig. 1 into a motion control IC in this study,
which is shown in Fig. 2. Our proposed motion control IC has two IPs (Intellectual
Properties). One IP performs the functions of the motion trajectory by software. The other IP

Fuzzy Controllers Recent Advances in Theory and Applications 182
performs the functions of two axes position/speed/current controllers by hardware. As the
results, this two IP will parallel processing in FPGA, and the hardware/software co-design
technology in FPGA can make the motion controller of X-Y more compact, flexible, better
performance and less cost. Further, the X-Y table usually leads to the existence of
unmodelled dynamics and disturbances which often significantly deteriorate the system
performance during a machining process. Many studies attempt to improve the tracking
performance in a machining process (Lin et al., 2006; Wang and Lee, 1999). Lin et al. (2006)
adopts a recurrent-neural-network sliding-mode controller to improve the motion tracking
performance of the X-Y table. Wang and Lee (1999) integrate the cross-coupled control and
neural network techniques to achieve a high accuracy of the motion tracking in the linear
motor X-Y table. However, due to the complicate computation of the neural-network, the
algorithms of above two studies are realized in the PC-based control system.

Figure 1. Conventional motion control system for X-Y table

Figure 2. Proposed FPGA-based motion control system for X-Y table
In this chapter, a motion control IC for linear motor drive X-Y table based on FPGA (Field
programmable gate array) technology is presented and shown in Fig.3. Firstly, the
mathematical model of the X-Y table is defined. Secondly, an adaptive fuzzy controller

FPGA-Based Motion Control IC for Linear Motor Drive X-Y Table Using Adaptive Fuzzy Control 183
(AFC) is introduced and adopted in position loop of X-Y table to improve the motion
tracking performance under unmodelled uncertainty condition. Thirdly, in implementation,
an FPGA embedded by a Nios II processor is used to design the overall circuits of the
motion control IC which the scheme of position/speed/current control for two PMLSMs
(permanent magnetic linear synchronous motors) is realized by hardware in FPGA and the
motion trajectory algorithm for X-Y table is implemented by software using Nios II
embedded processor. To reduce the FPGA resource usage, an FSM (Finite state machine)
joined by a multiplier, an adder, a LUT (Look-up table), some comparators and registers is
used to model the overall AFC algorithm. And VHDL (VHSIC hardware description
language) is adopted to describe the FSM. Herein, Altera Stratix II EP2S60, which has 48,352
ALUTs (Adaptive Look-Up Tables), total 2,544,192 RAM bits, and a Nios II embedded
processor which has a 32-bit configurable CPU core, 16 M byte Flash memory, 1 M byte
SRAM and 16 M byte SDRAM, is used. Therefore, a fully digital motion controller can be
implemented by an FPGA using hardware/software co-design technology which will make
the motion controller of the X-Y table more compact, flexible and better performance.
Finally, an experimental system is set up to verify the performance of the proposed motion
control IC for linear motor drive X-Y table.
2. System description of X-Y table and motion controller design
The internal architecture of the proposed FPGA-based controller system for a linear motor
drive X-Y table is shown in Fig. 3, in which the motion trajectory is implemented by
software using Nios II embedded processor; the position, speed and current vector
controller for two PMLSMs are implemented by hardware in FPGA chip. The mathematical
modeling of PMLSM, AFC algorithm and motion trajectory planning are introduced as
follows:

Figure 3. The architecture of a motion controller system for linear motor drive X-Y table

Fuzzy Controllers Recent Advances in Theory and Applications 184
2.1. Mathematical model of the PMLSM drive
The dynamic model of a typical PMLSM can be described in the synchronous rotating
reference frame, as follows

1
v
q
d s
d p q d
d d d
L
di R
i x i
dt L L L
t
t
= + + (1)


1
v

q f
d s
p d q p q
q q q q
di
L R
x i i x
dt L L L L
t
t
t t
= + (2)
where vd, vq are the d and q axis voltages; id, iq, are the d and q axis currents, Rs is the phase
winding resistance; Ld, Lq are the d and q axis inductance;
p
x is the translator speed;
f
is the
permanent magnet flux linkage; t is the pole pitch. The developed electromagnetic thrust
force is given by

3
(( ) )
2
e d q d f q
F L L i i
t

t
= + (3)
The current control of a PMLSM drive is based on a vector control approach. That is, if we
control id to 0 in Fig.3, the PMLSM will be decoupled, so that control a PMLSM will become
easy as to control a DC linear motor. After simplification and considering the mechanical
load, the model of a PMLSM can be written as the following equations,

3
2
e f q t q
F i K i
t

t
= A (4)
with

3

2
t f
K
t

t
=
(5)
and the mechanical dynamic equation of PMLSM in x-axis table is

2
2
p p
e L m m
d x dx
F F M B
dt dt
= + (6)
where
e
F ,
t
K ,
m
M ,
m
B and
L
F represent the motor thrust force, the force constant, the total
mass of the moving element, the viscous friction coefficient and the external force,
respectively. In addition, the current loop of the PMLSM drive in Fig.3 includes PI
controller, coordinate transformations of Clark, Modified inverse Clark, Park, inverse
Park, SVPWM (Space Vector Pulse Width Muldulation), pulse signal detection of the
encoder etc. The coordination transformation of the PMLSM in Fig. 3 can be described in
synchronous rotating reference frame. Figure 4 is the coordination system in rotating
motor which includes stationary a-b-c frame, stationary o-| frame and synchronously

FPGA-Based Motion Control IC for Linear Motor Drive X-Y Table Using Adaptive Fuzzy Control 185
rotating d-q frame. Further, the formulations among three coordination systems are
presented as follows.
1. Clarke : stationary a-b-c frame to stationary o-| frame.

2 1 1
3 3 3
1 1
0
3 3
a
b
c
i
i
i
i
i
o
|
(
(
(
(
(
(
=
(
(

(
(

(
(

(7)
2. Modified
1
Clarke

: stationary o-| frame to stationary a-b-c frame.

3 1
2 2
3 1
2 2
1 0
a
b
c
v
v
v
v
v
|
o

( (
(
( (
=
(
( (
(

( (

(8)
3. Park : stationary o-| frame to rotating d-q frame.

cos sin
sin cos
d e e
q e e
i i
i i
o
|
u u
u u
( ( (
=
( ( (

( (

(9)
4.
1
Park

: rotating d-q frame to stationary o-| frame.

cos sin
sin cos
d e e
q e e
v v
v v
o
|
u u
u u
( ( (
=
( ( (
( (

(10)
where
e
u is the electrical angle.
In Fig. 3, two digital PI controllers are presented in the current loop of PMLSM. For the
example in d frame, the formulation is shown as follows.

*
( ) ( ) ( )
d d d
e k i k i k =
(11)

_ _
( ) ( )
p d p d d
v k k e k =
(12)

_ _ _
( ) ( 1) ( 1)
i d i d i d d
v k v k k e k = +
(13)

_ _
( ) ( ) ( )
d p d i d
v k v k v k = +
(14)
the
d
e is the error between current command and measured current. The
_ p d
k ,
_ i d
k are P
controller gain and I controller gain, respectively. The
_
( )
p d
v k ,
_
( )
i d
v k , ( )
d
v k are the output of
P controller only, I controller only and the PI controller, respectively. Similarity, the
formulation of PI controller in q frame is the same.

Fuzzy Controllers Recent Advances in Theory and Applications 186

Figure 4. Transformation between stationary axes and rotating axes
2.2. Adaptive fuzzy controller (AFC) in position control loop
The green dash rectangular area in Fig. 3 presents the architecture of an AFC. It consists of a
fuzzy controller, a reference model and a parameter adjusting mechanism. Detailed
description of these is as follows.
1. Fuzzy controller (FC):
In Fig.3, the tracking error and the change of the error, e, de are defined as

( ) ( ) ( )
m p
e k k k =
(15)

( ) ( ) ( 1) de k e k e k =
(16)
and e, de and uf are input and output variables of FC, respectively. Besides
m
represents
m
x
or
m
y , and
p
represents
p
x or
p
y . The design procedure of the FC is as follows:
a. Take the e and de as the input variables of the FC, and define their linguist variables as E
and dE. The linguist value of E and dE are {A0, A1, A2, A3, A4, A5, A6} and {B0, B1, B2, B3, B4,
B5, B6}, respectively. Each linguist value of E and dE is based on the symmetrical
triangular membership function which is shown in Fig.5. The symmetrical triangular
membership function are determined uniquely by three real numbers
1 2 3
s s , if one
fixes
1 3
( ) ( ) 0 f f = = and
2
( ) 1 f = . With respect to the universe of discourse of [-6.6],
the numbers for these linguistic values are selected as follows:

{ } { } { }
{ } { } { } { }
0 0 1 1 2 2
3 3 4 4 5 5 6 6
: 6, 6, 4 , : 6, 4, 2 , : 4, 2,0 ,
: 2,0, 2 , : 0, 2, 4 , : 2, 4,6 , : 4,6, 6
A B A B A B
A B A B A B A B
= = =
= = = =
(17)
b. Compute the membership degree of the e and de. Figure 5 shows that the only two
linguistic values are excited (resulting in a non-zero membership) in any input value,
and the membership degree is obtained by
: 3-axis stationary frame
: 2-axis stationary frame
: 2-axis rotating frame q d
| o
c b a
a
c

f
q
f
q
f
rotor
stator
S
f
S
f
e
e
e
e
e
u
o
d
q
|
b
a
c
d
f
d
f
rotor
stator
S
f
S
f
e
e
e
e
o
d
q
|
b

FPGA-Based Motion Control IC for Linear Motor Drive X-Y Table Using Adaptive Fuzzy Control 187

1
( ) and
2
i
i
A
e e
e
+

=
1
( ) 1 ( )
i i
A A
e e
+
= (18)
where
1
6 2 * ( 1)
i
e i
+
A + + . Similar results can be obtained in computing the membership
degree ( )
j
B
de .
c. Select the initial fuzzy control rules, such as,

i j j,i
IF is A and e is B THEN is c
f
e u A (19)
where i and j = 0~6, Ai and Bj are fuzzy number, and cj,i is a real number. The graph of the
fuzzy rule table and the fuzzification are shown in Fig. 5.
d. Construct the output of the fuzzy system uf(e,de) by using the singleton fuzzifier,
product-inference rule, and central average defuzzifier method. Although there are
total 49 fuzzy rules in Fig. 5 will be inferred, actually only 4 fuzzy rules can be
effectively excited to generate a non-zero output. Therefore, if an error e is located
between ei and ei+1, and an error change de is located between dej and dej+1, only four
linguistic values Ai, Ai+1, Bj, Bj+1 and corresponding consequent values cj,i, cj+1,i, cj,i+1, cj+1,i+1
can be excited, and the output of the fuzzy system can be inferred by the following
expression:

1 1
, 1 1
1 , , 1
[ ( ) * ( )]
( , ) *
( ) * ( )
n m
n m
j i
m n A B j i
n i m j
f j m n n m i
n i m j
A B
n i m j
c e de
u e de c d
e de


+ +
+ +
= =
+ +
= =
= =
= A

(20)
where
,
( ) * ( )
n m
n m A B
d e de A . And those
, i j
c are adjustable parameters. In addition, by using
(18), it is straightforward to obtain
1 1
,
1
j i
n m
n i m j
d
+ +
= =
=

in (20).
2. Reference model (RM):
Second order system is usually as the RM in the adaptive control system. Therefore, the
transfer function of the RM in Fig.3 can be expressed as

2
* 2 2
( )
( ) 2
m n
p n n
s
s s s
e
e e
=
+ +
(21)
where
n
e is natural frequency and is damping ratio. Furthermore, because the
characteristics of no overshoot, fast response and zero steady-state error are the important
factors in the design of a PMLSM servo system; therefore, it can be considered as the
selective criterion of
n
e and . The design methodology is described as follows: Firstly, the
(21) matches the requirement of a zero steady-state error condition. Secondly, if we choose
1 > , it can guarantee no overshoot condition. Especially, the critical damp value 1 = has

Fuzzy Controllers Recent Advances in Theory and Applications 188
a fastest step response. Hence, the relation between the rising time tr and the natural
frequency
n
e for a step input response in (21) can be derived and shown as follows.

(1 ) 0.1
n r
t
n r
t e
e
e

+ =
(22)
Once the tr is chosen, the natural frequency
n
e can be obtained. Furthermore, applying the
bilinear transformation, (21) can be transformed to a discrete model by

1 1 2
0 1 2
* 1 1 2
1 2
( )
( ) 1
m
p
z a a z a z
z b z b z

+ +
=
+ +
(23)
and the difference equation is written as.

* * *
1 2 0 1 2
( ) ( 1) ( 2) ( ) ( 1) ( 2)
m m m p p p
k b k b k a k a k a k = + + +
(24)

Figure 5. The symmetrical triangular membership function of e and de, fuzzy rule table, fuzzy inference
and fuzzification
3. Parameter adjusting mechanism:
The gradient descent method is used to derive the AFC control law in Fig. 3. The objective of
the parameters adjustment in FC is to minimize the square error between the mover position
and the output of the RM. The instantaneous cost function is defined by

2
2
1 1
( 1) ( 1) ( 1) ( 1)
2 2
m m p
J k e k k k
(
+ = + = + +

(25)
and the four defuzzifier parameters of cj,i, cj+1,i, cj,i+1, cj+1,i+1 are adjusted according to

,
, ,
( 1) ( 1)
( 1)
( ) ( )
m n
m n m n
J k J k
c k
c k c k
o
c + c +
A + =
c c
(26)
2
e i 2 4
2
e e
e
1 i
A
i
+
=

=
+
*
) (
) ( ) ( e 1 e
i 1 i
A A
=
+
If is located between the and
then
e
i
e
1 i
e
+
c
00
dE
1
A0 A1
A2
A3
A
4
A
5
A
6
0 6 -6
e
e
E
d
e
d
e
(e)

(
d
e
)
1

A
3
(e)
A
4
(e)=1- A
3
(e)

B
1
(
d
e
)

B
2
(
d
e
)
=
1
-

B
1
(
d
e
)
Rule 1: e is A3 and de is B1 then u
f
is c13
Rule 2: e is A3 and de is B2 then u
f
is c23
Rule 3: e is A4 and de is B1 then u
f
is c14
Rule 4: e is A4 and de is B2 then u
f
is c24
Defuzzification
42 24 41 14 32 23 31 13
42 41 32 31
42 24 41 14 32 23 31 13
f
d c d c d c d c
d d d d
d c d c d c d c
u
* * * *
* * * *
+ + + =
+ + +
+ + +
=
where
0
6
-
6
Fuzzy Rule Table
Input of e (for i=3)
I
n
p
u
t

o
f

d
e


(
f
o
r

j
=
1
)
Fuzzy Inference and Output
A0 A1
A2 A
3
A4 A5 A6
c
01
c
02
c
03
c
04
c
05
c
06
c
10
c
11
c
12
c
13
c
14
c
15
c
16
c
20
c
21
c
22
c
23
c
24
c
25
c
26
c
30
c
31
c
32
c
33
c
34
c
35
c
36
c
40
c
41
c
42
c
43
c
44
c
45
c
46
c
50
c
51
c
52
c
53
c
54
c
55
c
56
c
60
c
61
c
62
c
63
c
64
c
65
c
66
B0
B1
B2
B3
B4
B5
B6
B
0
B
1
B
2
B
3
B
4
B
5
B
6
d
31
=
A3
(e)*
B1
(de)
d41 =A4
(e)*B1
(de) =(1-A3
(e))*B1
(de)
d
32
=
A3
(e)*
B2
(de)=
A3
(e)*(1-
B1
(de))
d
42
=
A4
(e)*
B2
(de)=(1-
A3
(e))*(1-
B1
(de))
4 2 -4 -2
4
2
-
4
-
2
and d31 +d
32
+d
41
+d
42
=1
1
(e)
Ai Ai+1
e
e
ei = -6+2*i e
i+1
= -4+2*i
) (e
i A

) (e
1 i A+

FPGA-Based Motion Control IC for Linear Motor Drive X-Y Table Using Adaptive Fuzzy Control 189
with m = j, j+1, n = i,i+1 and where o represents learning rate. However, following the
similar derivation with (Kung & Tsai, 2007), the
, m n
c A can be obtained as

, ,
( ) ( ) ( )
m n p i v n m
c k K K K e k d o A ~ + (27)
with m = j, j+1 and n = i,i+1.
2.3. Motion trajectory planning of X-Y table
The circular, window and star motion trajectories are typical used as the performance
evaluation of the motion controller for X-Y table.
a. In circular motion trajectory, it is computed by
sin( )
i i
x r u = (28)
os( )
i i
y r c u = (29)
with
1 i i
u u u

= + A . Where u A , r ,
i
x ,
i
y are angle increment, radius, X-axis trajectory
command and Y-axis trajectory command, respectively.
b. The window motion trajectory is shown in Fig.6. The formulation is derived as follows:
a-trajectory:

1 1
,
i i i i
x x y S y

= = + (30)
b-trajectory:
1
6
( : 2 , and )
4
i i i
u t t u u u

= + A

1 1
cos( ), sin( )
i x i i y i
x O r y O r u u = + = + (31)
c-trajectory:

1 1
,
i i i i
x S x y y

= + = (32)
d-trajectory:
1
6
( : , and )
4
i i i
u t t u u u

= + A

2 2
cos( ), sin( )
i x i i y i
x O r y O r u u = + = + (33)
e-trajectory:

1 1
,
i i i i
x x y S y

= = + (34)

Fuzzy Controllers Recent Advances in Theory and Applications 190
f-trajectory:

1
1
( : , and )
2
i i i
u t t u u u

= + A

3 3
cos( ), sin( )
i x i i y i
x O r y O r u u = + = + (35)
g-trajectory:

1 1
,
i i i i
x S x y y

= + = (36)
h-trajectory:
1
1
( : 0 , and )
2
i i i
u t u u u

= + A

4 4
cos( ), sin( )
i x i i y i
x O r y O r u u = + = + (37)
i-trajectory:

1 1
,
i i i i
x x y S y

= = + (38)
where S , u A ,
i
x ,
i
y are position increment, angle increment, X-axis trajectory command
and Y-axis trajectory command, respectively. In addition, the
1 1
( , )
x y
O O ,
2 2
( , )
x y
O O ,
3 3
( , )
x y
O O
,
4 4
( , )
x y
O O are arc center of b-, d-, f-, and h-trajectory in the Fig. 6 and r is the radius. The
motion speed of the table is determined by u A .

Figure 6. Window motion trajectory
c. Star motion trajectory is shown in Fig.7. The formulation is derived as follows:
a-trajectory :

1 1
,
i i i i
x S x y y

= + = (39)
1 1
( , )
x y
O O
x
y
2 2
( , )
x y
O O
4 4
( , )
x y
O O 3 3
( , )
x y
O O
a
b
e
f
c
d
Start
g
h
i
u A
1 i i
u u u

= + A
2 2
( , )
x y
O O
( , )
i i
x y
1 1
( , )
i i
x y
+ +
1 i
u

r

FPGA-Based Motion Control IC for Linear Motor Drive X-Y Table Using Adaptive Fuzzy Control 191
b-trajectory :

1 1
* sin54 , * sin36
o o
i i i i
x S x y S y

= + = + (40)
c-trajectory :

1 1
* sin18 , * sin72
o o
i i i i
x S x y S y

= + = + (41)
d-trajectory :

1 1
* sin18 , * sin72
o o
i i i i
x S x y S y

= + = + (42)
e-trajectory :

1 1
* sin54 , * sin36
o o
i i i i
x S x y S y

= + = + (43)
Where S ,
i
x ,
i
y are position increment, X-axis trajectory command and Y-axis trajectory
command, respectively. The motion speed of the table is determined byS .

Figure 7. Star motion trajectory
3. The design of a motion control IC for linear motor drive X-Y table
Figure 8 illustrates the internal architecture of the proposed FPGA-based motion control IC
for linear motor drive X-Y table. The FPGA uses Altera Stratix II EP2S60, which has 48,352
ALUTs (Adaptive Look-Up Tables), 36 DSP blocks, 144 embedded multipliers, 718
maximum user I/O pins, total 2,544,192 RAM bits, and a Nios II embedded processor which
has a 32-bit configurable CPU core, 16 M byte flash memory, 1 M byte SRAM and 16 M byte
SDRAM. The Nios II processor can be downloaded into FPGA to construct a SoPC
environment. The internal circuit in Fig. 8 comprises a Nios II embedded processor IP
(Intellectual Properties) and an application IP. The Nios II processor is depicted to both
generate the motion trajectory and collect the response data. The application IP includes the
y
x
b
e
a
c
d
Start
72

18

36

54

Fuzzy Controllers Recent Advances in Theory and Applications 192
circuits of two position AFC and speed P controllers as well as two current vector
controllers for X-axis and Y-axis table. The sampling frequency of position control loop is
designed with 2kHz. The operating clock rate of the designed FPGA controller is 50MHz
and the frequency divider generates 50 Mhz (Clk), 25 MHz (Clk-step), 2 kHz (Clk-sp) and 16
kHz (Clk-cur) clock to supply all module circuits of application IP in Fig. 8.
An FSM is also employed to model the AFC of the position loop and P controller of the speed
loop in X-axis table and shown in Fig. 9, which uses one adder, one multiplier, a look-up
table, comparators, registers, etc. and manipulates 35 steps machine to carry out the overall
computation. With exception of the data type in reference model are 24-bits, others data type
are designed with 12-bits length, 2s complement and Q11 format. Although the algorithm of
AFC is highly complexity, the FSM can give a very adequate modeling and easily be
described by VHDL. Furthermore, steps s0~s6 execute the computation of reference model
output; steps s6~s9 are for the computation of mover velocity, position error and error change;
steps s9~s12 execute the function of the fuzzification; s13 describe the look-up table and s14~s22
defuzzification; and steps s23~s34 execute the computation of velocity and current command
output, and the tuning of fuzzy rule parameters. The SD is the section determination of e and
de and the RS,1 represents the right shift function with one bit. The operation of each step in
Fig.9 can be completed within 40ns (25 MHz clock) in FPGA; therefore total 35 steps need a
1.4s operation time. It doesnt loss any control performance for the overall system because
the operation time with 1.4s is much less than the sampling interval, 500 s (2 kHz), of the
position control loop in Fig.3. In Fig. 8, the QEP circuit and circuit for current vector control
refer to (Kung & Tsai, 2007). Further, the Nios II embedded processor IP is depicted to
perform the function of the motion trajectory and two-axis position/speed loop controller for
X-Y table in software. Figure 10 illustrates the flow charts of the main program and the
interrupt service routine (ISR), where the interrupt interval is designed with 2ms. All
programs are coded in the C programming language. Then, through the complier and linker
operation in the Nios II IDE (Integrated Development Environment), the execution code is
produced and can be downloaded to the external Flash or SDRAM via JTAG interface.
Under the proposed design method, the overall resource usage of the proposed motion
control IC is listed in Table 1 which the two AFC circuits need 16,110 ALUTs, the Nios II
embedded processor IP needs 8,275 ALUTs and 46,848 RAM bits and the application IP
needs 22,928 ALUTs and 595,968 RAM bits in FPGA. Therefore, the motion control IC uses
64.5% ALUTs resource and 25.2% RAM resource of Stratix II EP2S60.
Table 1. The resource usage of a motion control IC in FPGA
Application IP
642,816 31,203
595,968 6,818
2 x Current loop controller
(Current vector control,
SVPWM,ADC,QEP)
0 16,110
2 x Adaptive fuzzy controller
(AFC)
Total
46,848 8,275 Nios II Embedded Processor IP
Memory
(bits)
ALUTs Module circuit IP
Application IP
642,816 31,203
595,968 6,818
2 x Current loop controller
(Current vector control,
SVPWM,ADC,QEP)
0 16,110
2 x Adaptive fuzzy controller
(AFC)
Total
46,848 8,275 Nios II Embedded Processor IP
Memory
(bits)
ALUTs Module circuit IP

FPGA-Based Motion Control IC for Linear Motor Drive X-Y Table Using Adaptive Fuzzy Control 193

Figure 8. The internal architecture of a motion control IC for linear motor drive X-Y table

Figure 9. State diagram of an FSM for describing the AFC in position loop and P controller in speed
loop (for X-axis Table)
Frequency
divider
QEP detection
and
transformation
CLK
CPU
On-chip
ROM
On-chip
RAM
UART
PIO
Timer
SPI
A
v
a
l
o
n

B
u
s
A
v
a
l
o
n

B
u
s
Nios II Embedded Processor IP
Altera FPGA (Stratix II EP2S60F672C5ES )
sram_cs
sram_we
sram_oe
sram_be[3]
sram_be[2]
sram_be[1]
sram_be[0]
D[31]
D[0]
A[22]
A[0]
Clk
Clk-ctr
Clk-sp
[15..0]
[15..0]
Application
IP
X-axis
Position
controller
STSB
RCB
RCA
STSA
CHB
CHA
xPWM 1
xPWM 2
xPWM 3
xPWM 4
xPWM 5
xPWM 6
xADIN[11]
xADIN[0]
xBDIN[11]
xBDIN[0]
Circuit of
current vector
controller
Clk
Clk-cur
Clk-step
xEncoder-A
xEncoder-B
xEncoder-Z
From Linear
Encoder of x-axis
Clk
Circuit of position
adaptive fuzzy
controller (AFC)
and speed P
controller
[15..0]
p
x
Clk
Clk-sp
*
p
x [15..0]
Clk-step
[11..0]
*
q
i
*
p
x
STSB
RCB
RCA
STSA
CHB
CHA
yPWM 1
yPWM 2
yPWM 3
yPWM 4
yPWM 5
yPWM 6
yADIN[11]
yADIN[0]
yBDIN[11]
yBDIN[0]
Circuit of
current vector
controller
Clk
Clk-cur
Clk-step
QEP detection
and
transformation
yEncoder-A
yEncoder-B
yEncoder-Z
From Linear
Encoder of y-axis
Clk
Circuit of position
adaptive fuzzy
controller (AFC)
and speed P
controller
[15..0]
p
y
Clk
Clk-sp
*
p
y [15..0]
Clk-step
[11..0] *
q
i
*
p
y
Y-axis
Position
controller
e
u [11..0]
e
u [11..0]
e
u [11..0]
e
u [11..0]
Clk-step
+
-
x
x
+
x
+
x
s
0
s
1
s
2
s
3
s
4
s
5
x
+
-
+
-
s
6
+
-
+
-
s
7
s
8
) (
*
k x
p
) ( 1 k x
p

) ( 1 k x
p

) ( 1 k x
m

) ( 1 k x
m

) (k v
) (k e
) 1 ( k e
) 1 ( k e
0
a
1
a
2
a
1
b
2
b
s
9
Computation of reference model output
Computation of mover velocity,
position error and error change
s
14
x
+
1
-
) (e
i A

) (de
j B

j i
d
,
+
-
x
x
x
s
15
s
16
s
17
) (de
j B

) e (
i A 1 +

1
) de (
j B 1 +

j , i
d
1 +
1 + j , i
d
1 1 + + j , i
d
) (de
j B

) (e
i
A

s
18
x
x
+
x
x
+ + x +
u
i
x +
s
20
s
21
s
23
s
24
s
28
s
19
s
22
s
25
i j
c
,
i , j
c
1 +
1 + i , j
c
1 1 + + i , j
c
j i
d
,
j i
d
, 1 +
1 , 1 + + j i
d
1 , + j i
d
i
K
p
K
f
u
) (
*
k i
q
x +
i j
c
,
i , j
c
1 +
1 + i , j
c
1 1 + + i , j
c
i , j
c
1 +
1 + i , j
c
j i
d
,
j , i
d
1 +
1 1 + + j , i
d
+ x x

i
K
p
K ) (k e
i j
c
,
x +
1 , + j i
d
x +
x +
s
29
s
30
s
31
s
32
s
33
s
34
r A
r A
r A
r A
r A
u
i
1 1 + + i , j
c
Defuzzification
Computation of velocity and current command, and tuning of fuzzy rule parameters
+
-
+
-
&
i
j
j&i
s
10
s
11
s
12
s
13
) (k de
) (e
i A

) (de
j B

i j
c
,
1 + i , j
c
i , j
c
1 +
1 1 + + i , j
c
1 + i
e
1 + j
de
Look-up
Fuzzy rule
table
RS,1
RS,1
Fuzzification
SD
k
e
SD
k
de
) (k e
) (
*
1 k x
p

) (
*
2 k x
p

) ( 2 k x
m

) (k x
m
) ( 1 k x
m
) ( 2 k x
m

) (
*
k x
p
) (
*
1 k x
p

) (
*
1 k x
p
) (
*
2 k x
p

) (k x
m
) (k x
p
) (k x
p
+
x
v
K
) (k v
x
v
K
-
s
26
s
27
) (k u
Look-up fuzzy table

Fuzzy Controllers Recent Advances in Theory and Applications 194

Figure 10. Flow chart of the main and ISR program in Nios II embedded processor
4. Experimental results
The overall experimental system depicted in Fig.3 includes an FPGA (Stratix II
EP2S60F672C5) experimental board, two voltage source IGBT inverters and an X-Y table
which is driven by two PMLSMs. The PMLSM was manufactured by the BALDOR electric
company; and it is a single-axis stage with a cog-free linear motor and a stroke length with
600mm. The parameters of the motor are: Rs = 27 O , Ld = Lq = 23.3 mH, Kt = 79.9N/A. The
input voltage, continuous current, peak current (10% duty) and continuous power of the
PMLSM are 220V, 1.6A, 4.8A and 54W, respectively. The maximum speed and acceleration
are 4m/s and 4 g but depend on external load. The moving mass is 2.5Kg, the maximum
payload is 22.5Kg and the maximum thrust force is 73N under continuous operating
conditions. A linear encoder with a resolution of 5m is mounted on the PMLSM as the
position sensor, and the pole pitch is 30.5mm (about 6100 pulses). The inverter has three sets
of IGBT power transistors. The collector-emitter voltage of the IGBT is rated 600V; the gate-
emitter voltage is rated 20V, and the DC collector current is rated 25A and in short time
(1ms) is 50A. The photo-IC, Toshiba TLP250, is used in the gate driving circuit of IGBT.
Input signals of the inverter are PWM signals from the FPGA device.
To confirm the effectiveness of the proposed AFC in linear drive X-Y table, a realization of
position controller based on the FPGA in Fig.3 is constructed and some experiments are
evaluated. The control sampling frequency of the current, speed and position loops are
designed as 16kHz, 2kHz and 2kHz, respectively. In the motion control IC, two
position/speed/current controllers are all realized by hardware in FPGA, and the motion
trajectory algorithm is implemented by software using the Nios II embedded processor. The
speed controller adopts a P controller and the AFC is used in the position loop. The transfer
function of the reference model is selected by a second order system with the natural
frequency of 30 rad/s and damping ratio of 1. The step response is first tested to evaluate the
performance of the proposed controller. Figures 11 and 12 respectively show the position
step responses for X-axis and Y-axis table using the FC (learning rate=0) and AFC (learning
Start of
main program
Initial interrupt
Initial timer
loop
Start of ISR
( each 500Hz )
Return
Output the X-axis and
Y-axis position command
Set parameters
of each axis
Computation of
position value for each axis
from motion trajectory
command

FPGA-Based Motion Control IC for Linear Motor Drive X-Y Table Using Adaptive Fuzzy Control 195
rate=0.1). The position command is a 4/3Hz square wave signal with 10mm amplitude. In
Figs. 11(a) and 12(a), when an 11 kg load is added upon the mover of the X-Y table and the
fuzzy control by using a fixed rule table, the position dynamic response in X-axis and Y-axis
table exhibits a 12.8% and 23.1% overshoot and severe oscillation, respectively. Accordingly,
an AFC is adopted in Fig.3. When the proposed AFC is used with learning rate being 0.1, the
tracking results are highly improved and presented in Figs. 11(c) and 12(c). Initially, the
position response in X-axis or Y-axis table tracks the output of the reference model with
oscillation. After one or two square wave commands, the ci,j parameters in fuzzy rule table
are tuned to adequate values, and the position response in X-axis or Y-axis table can closely
follow the output of the reference model. Further, the tracking motion about circular,
window and star trajectory by using FC and AFC are experimented. To evaluate the
tracking performance, the indices are firstly defined as follows.

2 2
( ) ( ( ) ( )) ( ( ) ( ))
m p m p
T k x k x x y k y x = +
(44)

1
( ) /
n
k
m T k n
=
=

(45)

2
1
( ( ) ) /
n
k
T k m n o
=
=

(46)
Where T(k), m and o respectively represent instantaneous value, mean and variance of
tracking error. In the circular tracking motion, the circle command is with center (25, 25) cm
and radius 10cm and its experimental results are shown in Figs. 13~14. In the window
tracking motion, the trajectory is designed as Fig.6 and its experimental results are shown in
Figs. 15~16. In the star tracking motion, the trajectory is designed as Fig.7 and its
experimental results are shown in Figs. 17~18. Further, the tracking performance in Figs
13~18 by using FC and AFC control algorithm are evaluated according to the indices of
(44)~(46), and its results are listed in Table 2. Compared with FC, the mean of tracking errors

Figure 11. (a) Position step response and (b) current response by using the FC as well as
(c) position step response and (d) current response by using the AFC in X-axis table
0.5 1 1.5 2 2.5 3 3.5 4
-5
0
5
10
15
Time (s)
P
o
s
i
t
i
o
n

r
e
s
p
o
n
s
e
(
m
m
)

0 0.5 1 1.5 2 2.5 3 3.5 4
-10
-5
0
5
10
Time (s)
I
q

c
u
r
r
e
n
t

(
A
)
Position command
Position response
Reference command
0
Using FC
(a)
(b)
0.5 1 1.5 2 2.5 3 3.5 4
-5
0
5
10
15
Time (s)
P
o
s
i
t
i
o
n

r
e
s
p
o
n
s
e
(
m
m
)

0 0.5 1 1.5 2 2.5 3 3.5 4
-10
-5
0
5
10
Time (s)
I
q

c
u
r
r
e
n
t

(
A
)
Position command Position response
Reference command
0
Using AFC
(c)
(d)

Fuzzy Controllers Recent Advances in Theory and Applications 196

Figure 12. (a) Position step response and (b) current response by using the FC as well as
(c) position step response and (d) current response by using the AFC in Y-axis table
in circular, winddow and star motion trajectory are significantly reduced about 41.6%, 14.6%
and 12.8% and the variance of tracking errors reduced about 33.3%, 64.6% and 47.4% after
using AFC. Therefore, it shows that the AFC has a better tracking performance than FC in
motion control of linear motor drive X-Y table. Finally, from the experimental results of
Figs.11~18, it demonstrates that the proposed AFC and the FPGA-based motion control IC
used for the linear motor drive X-Y table is effective and correct.
5. Conclusion
This study successfully presents a motion control IC for linear motor drive X-Y table based
on FPGA technology. The works herein are summarized as follows.
1. The functionalities required to build a fully digital motion controller of linear motor drive
X-Y table, such as the two current vector controllers, two speed P controllers, and two
position AFCs and one motion trajectory planning, have been integrated in one FPGA chip.
2. An FSM joined by one multiplier, one adder, one LUT, or some comparators and
registers has been employed to model the overall AFC algorithm, such that it not only is
easily implemented by VHDL but also the resources usage can be reduced in the FPGA.
3. The software/hardware co-design technology under SoPC environment has been
successfully applied to the motion controller of linear motor drive X-Y table.
However, the experimental results by step response as well as the circular, window and star
motion trajectory tracking, has been revealed that the software/hardware co-design technology
with the parallel processing well in the motion control system of linear motor drive X-Y table.

Table 2. Evaluation of tracking performance using FC and AFC
Variance
Mean
Tracking
error (mm)
Control
algorithm
1.85 0.75 0.40 3.52 2.12 0.60
1.43 1.22 2.05 1.64 1.43 3.51
Star
motion
Window
motion
Circular
motion
Star
motion
Window
motion
Circular
motion
Adaptive fuzzy controller
(AFC)
Fuzzy controller
(FC)
Variance
Mean
Tracking
error (mm)
Control
algorithm
1.85 0.75 0.40 3.52 2.12 0.60
1.43 1.22 2.05 1.64 1.43 3.51
Star
motion
Window
motion
Circular
motion
Star
motion
Window
motion
Circular
motion
Adaptive fuzzy controller
(AFC)
Fuzzy controller
(FC)
1 2 3 4 5 6 7 8
-5
0
5
10
15
Time (s)
P
o
s
i
t
i
o
n

r
e
s
p
o
n
s
e
(
m
m
)

0 1 2 3 4 5 6 7 8
-10
-5
0
5
10
Time (s)
I
q
c
u
r
r
e
n
t

(
A
)
Position command Position response
Reference command
Using FC
(a)
(b)
1 2 3 4 5 6 7 8
-5
0
5
10
15
Time (s)
P
o
s
i
t
i
o
n

r
e
s
p
o
n
s
e
(
m
m
)

0 1 2 3 4 5 6 7 8
-10
-5
0
5
10
Time (s)
I
q
c
u
r
r
e
n
t

(
A
)
Position command
Position response Reference command
Using AFC
(c)
(d)

FPGA-Based Motion Control IC for Linear Motor Drive X-Y Table Using Adaptive Fuzzy Control 197

Figure 13. Circular trajectory response by using the FC (a) Star trajectory tracking (b) Position tracking
in X- and Y-axis table (c) Control efforts in X- and Y-axis table (d) Tracking errors in X- and Y-axis table

Figure 14. Circular trajectory response by using the AFC (a) Star trajectory tracking (b) Position
tracking in X- and Y-axis table (c) Control efforts in X- and Y-axis table (d) Tracking errors
in X- and Y-axis table
150 200 250 300 350
140
160
180
200
220
240
260
280
300
320
340
360
X-axis Position (mm)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n

(
m
m
)

0 1 2 3 4 5
100
200
300
400
Time (s)
X
-
a
x
i
s

P
o
s
i
t
i
o
n
(
m
m
)

0 1 2 3 4 5
100
200
300
400
Time (s)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n
(
m
m
)

0 1 2 3 4 5
-4
-2
0
2
4
Time (s)
X
-
a
x
i
s

I
q

(
A
)

0 1 2 3 4 5
-4
-2
0
2
4
Time (s)
Y
-
a
x
i
s

I
q

(
A
)

0 1 2 3 4 5
-10
0
10
Time (s)
X
-
a
x
i
s

P
o
s
i
t
i
o
n

e
r
r
o
r
(
m
m
)

0 1 2 3 4 5
-10
0
10
Time (s)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n

e
r
r
o
r
(
m
m
)

(a)
(d)
(b)
(c)
-6.2
6.2
11.5
-5.7
4.6
150 200 250 300 350
150
200
250
300
350
X-axis Position (mm)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n

(
m
m
)

0 1 2 3 4 5
100
200
300
400
Time (s)
X
-
a
x
i
s

P
o
s
i
t
i
o
n
(
m
m
)

0 1 2 3 4 5
100
200
300
400
Time (s)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n
(
m
m
)

0 1 2 3 4 5
-4
-2
0
2
4
Time (s)
X
-
a
x
i
s

I
q

(
A
)

0 1 2 3 4 5
-5
0
5
Time (s)
Y
-
a
x
i
s

I
q

(
A
)

0 1 2 3 4 5
-10
0
10
Time (s)
X
-
a
x
i
s

P
o
s
i
t
i
o
n

e
r
r
o
r
(
m
m
)

0 1 2 3 4 5
-10
0
10
Time (s)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n

e
r
r
o
r
(
m
m
)

(a)
(d)
(b)
(c)
3.0
3.5
-2.7
-3.0

Fuzzy Controllers Recent Advances in Theory and Applications 198

Figure 15. Window trajectory response by using the FC (a) Star trajectory tracking (b) Position tracking
in X- and Y-axis table (c) Control efforts in X- and Y-axis table (d) Tracking errors in X- and Y-axis table

Figure 16. Window trajectory response by using the AFC (a) Star trajectory tracking (b) Position tracking
in X- and Y-axis table (c) Control efforts in X- and Y-axis table (d) Tracking errors in X- and Y-axis table
50 100 150 200 250
100
150
200
250
300
X-axis Position (mm)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n

(
m
m
)

0 2 4 6 8 10 12
0
100
200
300
Time (s)
X
-
a
x
i
s

P
o
s
i
t
i
o
n
(
m
m
)

0 2 4 6 8 10 12
100
200
300
Time (s)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n
(
m
m
)

0 2 4 6 8 10 12
-5
0
5
Time (s)
X
-
a
x
i
s

I
q

(
A
)

0 2 4 6 8 10 12
-4
-2
0
2
4
Time (s)
Y
-
a
x
i
s

I
q

(
A
)

0 2 4 6 8 10 12
-10
0
10
Time (s)
X
-
a
x
i
s

P
o
s
i
t
i
o
n

e
r
r
o
r
(
m
m
)

0 2 4 6 8 10 12
-10
0
10
Time (s)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n

e
r
r
o
r
(
m
m
)

(a)
(d)
(b)
(c)
3.1
-3.1
7.7
-8.0
50 100 150 200 250
100
150
200
250
300
X-axis Position (mm)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n

(
m
m
)

0 2 4 6 8 10 12
0
100
200
300
Time (s)
X
-
a
x
i
s

P
o
s
i
t
i
o
n
(
m
m
)

0 2 4 6 8 10 12
100
200
300
Time (s)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n
(
m
m
)

0 2 4 6 8 10 12
-5
0
5
Time (s)
X
-
a
x
i
s

I
q

(
A
)

0 2 4 6 8 10 12
-5
0
5
Time (s)
Y
-
a
x
i
s

I
q

(
A
)

0 2 4 6 8 10 12
-10
0
10
Time (s)
X
-
a
x
i
s

P
o
s
i
t
i
o
n

e
r
r
o
r
(
m
m
)

0 2 4 6 8 10 12
-10
0
10
Time (s)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n

e
r
r
o
r
(
m
m
)

(a)
(d)
(b)
(c)
5.0
-5.0
-1.9
1.8

FPGA-Based Motion Control IC for Linear Motor Drive X-Y Table Using Adaptive Fuzzy Control 199

Figure 17. Star trajectory response by using the FC (a) Star trajectory tracking (b) Position tracking in
X- and Y-axis table (c) Control efforts in X- and Y-axis table (d) Tracking errors in X- and Y-axis table

Figure 18. Star trajectory response by using the AFC (a) Star trajectory tracking (b) Position tracking in
X- and Y-axis table (c) Control efforts in X- and Y-axis table (d) Tracking errors in X- and Y-axis table
100 150 200 250 300
0
20
40
60
80
100
120
140
160
180
200
220
X-axis position (mm)
Y
-
a
x
i
s

p
o
s
i
t
i
o
n

(
m
m
)

0 5 10 15
100
200
300
400
Time (s)
X
-
a
x
i
s

t
r
a
c
k
i
n
g

(
m
m
)

0 5 10 15
0
50
100
150
200
250
Time (s)
Y
-
a
x
i
s

t
r
a
c
k
i
n
g

(
m
m
)

0 5 10 15
-5
0
5
Time (s)
X
-
a
x
i
s

I
q
c
u
r
r
e
n
t

(
A
)

0 5 10 15
-5
0
5
Time (s)
Y
-
a
x
i
s

I
q
c
u
r
r
e
n
t

(
A
)

0 5 10 15
-10
0
10
Time (s)
X
-
a
x
i
s

t
r
a
c
k
i
n
g

e
r
r
o
r
(
m
m
)

0 5 10 15
-10
0
10
Time (s)
Y
-
a
x
i
s

t
r
a
c
k
i
n
g

e
r
r
o
r
(
m
m
)

(a)
(d)
(b)
(c)
350
9.0
-11.4
-3.2
3.5
100 150 200 250 300
0
20
40
60
80
100
120
140
160
180
200
220
X-axis Position (mm)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n

(
m
m
)

0 5 10 15
100
200
300
400
Time (s)
X
-
a
x
i
s

P
o
s
i
t
i
o
n
(
m
m
)

0 5 10 15
0
50
100
150
200
250
Time (s)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n
(
m
m
)

0 5 10 15
-5
0
5
Time (s)
X
-
a
x
i
s

I
q

(
A
)

0 5 10 15
-5
0
5
Time (s)
Y
-
a
x
i
s

I
q

(
A
)

0 5 10 15
-10
0
10
Time (s)
X
-
a
x
i
s

P
o
s
i
t
i
o
n

e
r
r
o
r
(
m
m
)

0 5 10 15
-10
0
10
Time (s)
Y
-
a
x
i
s

P
o
s
i
t
i
o
n

e
r
r
o
r
(
m
m
)

(a)
(d)
(b)
(c)
350
-2.3
2.2
8.0
-10

Fuzzy Controllers Recent Advances in Theory and Applications 200
Author details
Ying-Shieh Kung
Department of Electrical Engineering, Southern Taiwan University, Taiwan
Chung-Chun Huang and Liang-Chiao Huang
Green Energy and Environment Research Laboratories, Industrial Technology Research Institute,
Taiwan
Acknowledgement
The financial support provided by Bureau of Energy is gratefully acknowledged.
6. References
Cho, J. U., Le, Q. N. and Jeon, J. W. (2009) An FPGA-based multiple-axis motion control
chip, IEEE Trans. Ind. Electron., vol. 56, no.3, pp.856-870.
Groove, M.P. (1996) Fundamentals of modern manufacturing: materials, process, and systems.
Prentice Hall, Upper Saddle River, NJ.
Goto, S., Nakamura, M. and Kyura, N. (1996) Accurate contour control of mechatronic servo
systems using gaussian networks. IEEE Trans. Ind. Electron., vol.43, no. 4, pp. 469-476.
Hanafi, D., Tordon, M. and Katupitiya, J. (2003) An active axis control system for a
conventional CNC machine. Proceedings of IEEE/ASME International Conference on
Advanced Intelligent Mechatronics, pp. 1188-1193.
Hall, T.S., Hamblen, J.O. (2004) System-on-a-programmable -chip development platforms in
the classroom. IEEE Trans. on Education, vol.47, no.4, pp.502-507.
Lin, F.J., SHieh, P.H. and Shen, P.H. (2006) Robust recurrent-neural-network sliding-mode
control for the X-Y table of a CNC machine. IEE Proc. Control Theory Application, vol.
153, no. 1, pp. 111-123.
Monmasson, E., Idkhajine L. and Naouar, M.W. (2011) FPGA-based Controllers. IEEE Trans.
Electron. Magaz., vol. 5, no.1, pp.14-26.
Kung, Y.S., Huang P.G. and Chen, C.W. (2004) Development of a SOPC for PMSM drives.
Proceedings of the IEEE International Midwest Symposium on Circuits and Systems, vol. II,
pp. II-329~II-332.
Kung, Y.S. and Tsai, M.H. (2007) FPGA-based speed control IC for PMSM drive with
adaptive fuzzy control. IEEE Trans. on Power Electronics, vol. 22, no. 6, pp. 2476-2486.
Kung, Y.S. and Chen, C.S. (2008). Realization of a motion control IC for robot manipulator
based on novel FPGA technology. Robot manipulators, programming, design and control.
pp.291~312, I-Tech, Vienna.
Sanchez-Solano, S., Cabrera, A. J., Baturone, I., Moreno-Velo, F.J., Brox, M. (2007) FPGA
implementation of embedded fuzzy controllers for robotic applications. IEEE Trans. Ind.
Electron., vol. 54, no. 4, pp. 1937-1945.
SOPC World, (2004) Altera Corporation.
Wang, G.J. and Lee, T.J. (1999) Neural-network cross-coupled control system with
application on circular tracking of linear motor X-Y table. International Joint Conference
on Neural Networks, pp. 2194-2199.
Chapter 9
Novel Yinger Learning Variable
Universe Fuzzy Controller
Ping Zhang and Guodong Gao
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/48266
1. Introduction
Fuzzy control is a practical alternative for a variety of challenging of challenging control
applications because it provides a convenient method for constructing nonlinear controllers
via the use of heuristic information. The heuristic information may come from an operator
who has acted as a human controller for a process. In the fuzzy control design methodology,
a set of rules are written down by the operator on how to control the process, then make
these into a fuzzy controller that emulates the decision-making process of the human. In
some cases, the heuristic information may come from other novel intelligent applications. In
other cases, the heuristic information may come from a control engineer who has performed
extensive mathematical modeling, analysis, and development of control algorithms for a
particular process. Regardless of where the heuristic control knowledge comes from, fuzzy
control provides a user-friendly formalism for representing and implementing the ideas.
Over the past few decades, fuzzy logic theory is widely used: process control, management
and decision making, operations research, economies. Dealing with simple yes and no
answers is no longer satisfactory enough; a degree of membership (Zadeh, 1965) became a
new way of solving problems. Fuzzy logic derives from the truth that the human common
sense reasoning mode is approximate in nature.
In this chapter we provide a control engineering perspective on novel fuzzy controller. We
take a pragmatic engineering approach to the design, analysis, performance evaluation, and
implement of fuzzy control system. The chapter is basically broken into five parts. In section
2, we provide an overview of conventional control system design. In section 3 the basic
theories of variable universe fuzzy control are been introduced. In section 4, we cover the
novel fuzzy controller based on Yinger algorithm. In section 5, we use some examples to
show how to design, simulate, and implement these controllers. Finally, in section 6, we

Fuzzy Controllers Recent Advances in Theory and Applications 202
explain how to write a computer program to simulate the novel fuzzy control system, using
either a high-level language or Matlab.
2. Conventional control system design
2.1. Introduction
A control system is a device, or set of devices to manage, command, direct or regulate the
behavior of other devices or system. There are two common classes of control systems, with
many variations and combinations: logic or sequential controls, and feedback or linear
controls. There is also fuzzy logic, which attempts to combine some of the design simplicity
of logic with the utility of linear control. Some devices or systems are inherently not
controllable. A basic control system is shown in figure 1. The plant is object to be controlled.
Its inputs are ( ) u t , its outputs are ( ) y t , and reference input is ( ) r t .
( ) u t ( ) r t ( ) y t

Figure 1. Control system
2.2. Mathematical modeling
The mathematical model is a description of a system using mathematical concepts and
language. The process of developing a mathematical model is termed mathematical
modeling. Mathematical models are used not only in the natural sciences (such as physics,
biology, earth science, meteorology) and engineering disciplines (such as computer science,
artificial intelligence), but also in the social sciences (such as economics, psychology,
sociology and political science),physicists, engineers, statisticians, operations research
analysts and economists use mathematical models most extensively. A model may help to
explain a system and to study the effects of different components, and to make predictions
about behaviour.
Mathematical models can take many forms, including but not limited to dynamical systems,
statistical models, differential equations, or game theoretic models. These and other types of
models can overlap, with a given model involving a variety of abstract structures. In
general, mathematical models may include logical models, as far as logic is taken as a part of
mathematics. In many cases, the quality of a scientific field depends on how well the
mathematical models developed on the theoretical side agree with results of repeatable
experiments. Lack of agreement between theoretical mathematical models and experimental
measurements often leads to important advances as better theories are developed.

Novel Yinger Learning Variable Universe Fuzzy Controller 203
When a control engineer is given a control problem, often one of the first tasks is the
development of a mathematical model of the process to be controlled, in order to gain a
clear understanding of the problem. Basically, there are only a few ways to actually generate
the model. We can use first principles of physics to write down a model. Another way is to
perform "system identification" via the use of real plant data to produce a model of the
system. Sometimes a combined approach is used where we use physics to write down a
general different equation that we believe represent the plant behavior, and then we
perform experiments on the plant to determine certain model parameters or functions.
Often, more than one mathematical model is produced. A "truth model" is one that is
developed to be as accurate as possible so that it can be used in simulation-based
evaluations of control systems. It must be understood, however, that there is never a perfect
mathematical model for the plant. The mathematical model is an abstraction and hence
cannot perfectly represent all possible dynamics of any physical process. This is not to say
that we cannot produce models that are "accurate enough" to closely represent the behavior
of a physical system. Usually, control engineer to be able to design a controller that will
work. Then, they often also need a very accurate model to test the controller in simulation
before it is tested in an experimental setting. Hence, lower-order "design model" are also
often developed that may satisfy certain assumption yet still capture the essential plant
behavior. Indeed, it is quite an art to produce good low-order model that satisfy these
constraints. We emphasize that the reason we often need simpler models is that the
synthesis techniques for controller often require that the model of the plant satisfy certain
assumptions or there methods generally cannot be used.
Linear models such as the one in Equation (1) have been used extensively in the past and the
control theory for linear system is quite mature.

= +
= +

x Ax Bu
y Cx Du
(1)
In this case u is the m-dimensional input; x is the n-dimensional state; y is the p-dimensional
output; and A,B,C and D are matrices of appropriate dimension. Such models are
appropriate for use with frequency domain design techniques, the root-locus method, state-
space methods, and so on. Sometimes it is assumed that the parameters of the linear model
are constant but unknown, or can be perturbed form their nominal values.
Much of the current focus in control is on the development of controllers using nonlinear
models of the plant of the form

=
=

( , )
( , )
x f x u
y g x u
(2)
Where the variables are defined as for the linear model and f and g are nonlinear functions
of their arguments. One form of the nonlinear model that has received significant attention
is

Fuzzy Controllers Recent Advances in Theory and Applications 204
= +

( ) ( ) x f x g x u (3)
Since it is possible to exploit the structure of this model to construct nonlinear controllers. Of
particular with both of the above nonlinear models is the case where f and g are not
completely known and subsequent research focuses on robust control of nonlinear system.
Discrete time versions of the above models are also used, and stochastic effect are often
taken into account via the addition of a input or other stochastic effects. Under certain
assumptions you can linearize the nonlinear model in Equation(2) to obtain a linear one. In
this case we sometimes think of the nonlinear model as the truth model, and the linear
model that are generated form it as control design model.
There are certain properties of the plant that the control engineer often seeks to identify early
in the design process. For instance, the stability of the plant may be analyzed. The effects of
certain nonlinearities are also studied. The engineer may want to determine if the plant is
controllable to see, for example, if the control input will be able to properly affect the plant;
and observable to see, for example, if the chosen sensors will allow the controller observe the
critical plant behavior so that it can be compensated Overall, this analysis of the plant's
behavior gives the control engineer a fundamental understanding of the plant dynamics.
2.3. Performance objectives and design constrains
Controller design entails constructing a controller to meet the specifications. Often the first
issue to address is whether to use open or closed-loop control. Often, need to pay for a
sensor for the feedback information and there need to justification for this cost. Moreover,
feedback can destabilize the system. Do not develop a feedback controller just because you
are used to developing feedback controllers; you may want to consider an open-loop
controllers since it may provide adequate performance. Assuming you use feedback control,
the closed-loop specifications can involve the following factors: Disturbance rejection
properties; Insensitivity to plant parameter variations; Stability; Rise-time.
2.4. Controller design
Conventional control has provided numerous methods for controllers for dynamic system.
Some of there are listed below:
1. Proportional-integral-derivative(PID) control:Over 90% of the controllers in operation
today are PID controllers. This approach is often viewed as simple, reliable,and easy to
understand. Often, like fuzzy controller, heuristics are used to tune PID controllers.
2. State-space methods: State feedback,observers,and so on.
3. Optimal control: Linear quadratic regulator,use of Pontryagin's minimum principle or
dynamic programming,an so on.
4. Nonlinear methods: Feedback linearization, Lyapunov redesign, sliding mode control,
backstepping, and so on.
5. Adaptive control; model reference adaptive control,self-tuning regulators, nonlinear
adaptive control,and so on.

Novel Yinger Learning Variable Universe Fuzzy Controller 205
Basically,there conventional approaches to control system design offer a variety of ways to
utilize information from mathematical model on how to do good control. Sometimes they
do not take into account certain heuristic information early in the design process, but use
heuristics when the controller is implemented to tune it(tuning is invariably needed since
the model used for the controller development is not perfectly accurate).Unfortunately,
when using some approaches to conventional control, some engineers become somewhat
removed from the control problem, and sometimes this leads to the development of
unrealistic control laws. Sometimes in conventional control, useful heuristics are ignored
because they do not fit into the proper mathematical framework, and this can cause
problem.
2.5. Performance evaluation
The next step in the design process is to perform analysis and performance evaluation.
Basically, we need performance evaluation to test that we design does in fact meet the
closed-loop specifications. This can be particularly important in safety-critical applications
such as the control of a washing machine or an electric shaver, it may not be as important in
the sense that failures will not imply the loss of life, so some of the rigorous evaluation
methods can sometimes be ignored. Basically, there are three general ways to verify that a
control system is operating properly: (1) mathematical analysis based on the use of formal
models, (2) simulation-based analysis that most often uses formal models, and (3)
experimental investigations on the real system.
3. Variable fuzzy control system design
The fuzzy controller block diagram is given in figure 2. The plant outputs are denoted by
( ) y t , its input is denoted by ( ) u t , and the reference input to the fuzzy controller is denoted
by ( ) r t .
( ) u t
( ) r t ( ) y t

Figure 2. Fuzzy controller architecture
3.1. Fuzzy controller
Basically, the difficult task of modeling and simulating complex real-world systems for
controller systems development, especially when implementation issues are considered, is

Fuzzy Controllers Recent Advances in Theory and Applications 206
well documented. Even if a relatively accurate model of a dynamic system can be
developed, it is often too complex to use require restrictive assumptions for the plant. It is
for this reason that in practice conventional controllers are often developed via simple
models of the plant behavior that satisfy the necessary assumptions, and via the ad hoc
tuning of relatively simple linear or nonlinear controllers. Regardless, it is well understood.
Fuzzy control provides a formal methodology for representing, manipulating, and
implementing a humans heuristic knowledge about how to control a system.
The fuzzy controller block diagram is given in Figure 2, where we show a fuzzy controller
embedded in a closed-loop control system. The plant outputs are denoted by y(t), its inputs
are denoted by u(t), and the reference input to the fuzzy controller is denoted by r(t).
The fuzzy controller has four main components: (1) The rule-base holds the knowledge, in
the form of a set of rules are relevant at the current time and then decides what the input to
the plant should be, (3) The fuzzification interface simply modifies the inputs so that they
can be interpreted and compared to the rules in the rule-base. And (4) the defuzzification
interface converts the conclusions reached by the inference mechanism into the inputs to the
plant.
To design the fuzzy controller, the control engineer must gather information on how the
artificial decision maker should act in the closed-loop system. Sometimes this information
can come from a human decision maker who performs the control task, while at other times
the control engineer can come to understand the plant dynamics and write down a set of
rules about how to control the system without outside help. These rules basically say, If
should be some value. A whole set of such If-Then rules is loaded into the rule-base, and
specifications are met.
3.2. Structure of variable adaptive fuzzy controller
Let ( = =

, ( 1, 2, , )
i
X E E i n be the universe of input variable = ( 1, 2, , )
i
x i n , and
= [ , ] Y U U be the universe of output variable y .
s s
=
(1 )
{ }
i ij j m
A stands for a fuzzy partition
on
i
X ,and
s s
B =
(1 )
{ }
j j m
B defines a fuzzy partition onY . A group of fuzzy inference rules is
formed as follow:
If
1
x is
1j
A and
2
x is
2j
A andand
n
x is
nj
A then y is
j
B , = 1, 2, , j m
The fuzzy logic system can be represented as an n-dimension piecewise interpolation
function
2
( , , , )
n
F x x x :

= =
= =
[

2 2
1 1
( , , , ) ( , , , ) ( )
n m
n n ij i j
j i
F x x x y x x x A x y (4)
Generally speaking, a function o o : [0,1], ( ) X x x can be called a contraction-expansion
factor on ( =

,
i
X E E . The so-called variable universe means
i
X andY can change with
changing variable
i
x and y expressed by:

Novel Yinger Learning Variable Universe Fuzzy Controller 207
o o = ( ) [ ( ) , ( ) ]
i i i i i i
X x x E x E (5)
| | = ( ) [ ( ) , ( ) ] Y y y U y U (6)

Figure 3. Universe compress and expand
4. Novel fuzzy controller based on Yinger algorithm
Novel fuzzy controller is composed of three parts. Firstly, new kind of contraction-
expansion factor is established, then local space is optimized, finally novel controller
dynamically adjust output by rules.
4.1. Optimal local spaces
Many real-world environments in which learning systems have to operate are time-varying.
Several aspects of the learning problem can vary, including the mapping to be learned, and
the sampling distribution that governs the input-space location of exemplars that make up
the input information. In this section, K-Vector Nearest Neighbors (K-VNN) is proposed to
this problem.
Define 1. Lets O
k
is input sets which can be defined to local space as:

{ } { }
O = <
1
, , ( , )
k K i i m
X X X D X X h (7)
Where h is radius of local space ( O
k
), and data-window is changed by adjusting it.
( , ) D A B is the distance function which is defined by (8),
1
, ,
K
X X are messages to input.
Define 2. Lets =
1
[ , , ]
n
A A A and =
1
[ , , ]
n
B B B ,in the Euclidean space, gets distance and
intersection angle:

u

2
2 2
( , )
( , ) arccos
T
d A B A B
A B
A B
A B
(8)

Fuzzy Controllers Recent Advances in Theory and Applications 208
According to (7), we can get the distance and intersection angle of
i
X and
d
X , from input-
output specimen choice similar message to O
k
.
If intersection angle of
i
X and
d
X .greater than

90 , thinking
i
X stray from
d
X ,and define
as follows:

(

( = +

s s s s
( , )
( , ) sin ( , )
(0 1 / 2,0 1 / 2)
i d
d X X
i d i d
D X X ae b X X
a b
(9)
From (9), we can see, if
i
X is more similar to
d
X ,
(

( , )
i d
d X X
e and (

sin ( , )
i d
X X are more
similar to 1, use this method and get the new input set

{ }
O = > >
1 1 1
( , ), ,( , ) ( , ) ( , )
k K K d k d
X Y X Y D X X D X X (10)
From this section, some noise can be deleted by this section.
4.2. Contraction-expansion factor
Now the popular contraction-expansion factor is o

=
2
( )
( ) 1
kx
x ce ( e > (0 1) 0 c k ), but the
algorithm module can not be realized easily by C++ which support some methods by using
VC++ accomplish control system. So building up a kind of contraction-expansion factor to
nonlinear system is very important.
1. Establish differential equation
Firstly, o( ( )) e t is strictly monotonously increasing on [0 1] and monotonously decreasing on
[-1 0].
Secondly, ( ) 0 e t Then o c = ( ( )) 0.0001 e t and ( ) 1 e t then o ( ( )) 1 e t .
Thirdly, o A = A ( ( )) ( ) e t k e t , and to the same A ( ) e t , ( ) e t is larger and o A ( ( )) e t is larger too.
From those conditions the differential equation can be build as follow:
o A = A
2
( ( )) ( ) ( )( ( )) e t ke t e t E e t (11)
get hold of:
o = + +
4 2
1
( )
4 2
E
x kx kx c (12)
and initialized condition:
when = ( ) 0 e t theno = ( ( )) ( , )
i d
e t D X X ,and ( ) e t =E o = ( ( )) 1 e t
get hold of:
o = + +
4 2
,
1
( ) ( )
4 2
i d
E
x kx kx D X X (13)

Novel Yinger Learning Variable Universe Fuzzy Controller 209
2. Analyze and verify characters
1. Duality o o e = ( ) ( ( )) ( ( )) e t E e t e t
2. Near zero o = >
,
(0) ( ) 0
i d
D X X
3. Monotonicity: e
1 2
( ), ( ) [0, ] e t e t E if >
1 2
( ) ( ) e t e t then o o >
1 2
( ( )) ( ( )) e t e t
4. Normality o o = = ( ) ( ) 1 E E
o = + +
4 2
,
1
( ) ( )
4 2
i d
E
x kx kx D X X
is the primary function can easily realize in nonlinear system. So the new kind of
contraction-expansion factor is satisfied with the requests.

Figure 4. Function cluster surface (k>0)

Figure 5. Function cluster surface contour(k>0)
-1
-0.5
0
0.5
1
0
5
10
0
0.2
0.4
0.6
0.8
1

0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
0
1
2
3
4
5
6
7
8
9
10

Fuzzy Controllers Recent Advances in Theory and Applications 210

Figure 6. Function cluster surface (c<1)

Figure 7. Function cluster surface contour (c<1)
5. Examples
Choose the typical non-linear system to the new algorithm.
Plant:

= +
+
=
( ) .
( )
1
( ) ( )
1
( ) ( )
x t
x t
e
x t u t
e
y t x t
(14)


= = ( ) ( ) ( ) 0
lim lim
t t
e t r t y t
(15)
And = ( ) ( )
c
u t u t
-1
-0.5
0
0.5
1
0
0.1
0.2
0.3
0.4
0
0.2
0.4
0.6
0.8
1



0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4

Novel Yinger Learning Variable Universe Fuzzy Controller 211
[ 1,1] U

We can get the rules as follows:
if e is Nb then u is NB
,
if e is Nm then u is NM

if e is Ns then u is NS
,
if e is Pb then u is PB

if e is Pm then u is PM
,
if e is Ps then u is PS


Figure 8. Function

+
= s s

1 , 3
1.9
( ) , 3 1.9
1.1
0 ,
e
e
Nb e e
etc

+
s s

+
= s s

2.4
, 2.4 1.5
0.9
0.6
( ) , 1.5 0.6
0.9
0 ,
e
e
e
Nm e e
etc