A Laboratory Course On Fuzzy Control Education, Ieee Transactions On
IEEE TRANSACTIONS ON EDUCATION, VOL. 42, NO. 1, FEBRUARY 1999
15
A Laboratory Course on Fuzzy Control
Stephen Yurkovich, Senior Member, IEEE, and Kevin M. Passino, Senior Member, IEEE
Abstract—In this paper we describe a new control laboratory
With these thoughts in mind, it has been the goal of the
course at The Ohio State University. Students execute a series
Ohio State control group to bring the newest technologies into
of laboratory exercises, for a variety of processes, implementing
the curricula, through both lecture and laboratory courses. In
fuzzy control, adaptive fuzzy control, and other intelligent control
1993, the authors were awarded a National Science Founda-
techniques, with a particular focus on fuzzy control. Fully instru-
mented independent testbeds in the laboratory emphasize several
tion Combined Research-Curriculum Development grant for
sensing and actuation technologies. Both senior undergraduates
incorporating new techniques in intelligent control into the
and graduate students take the course and several have used the
curriculum. The incorporation of existing research results in
experience, coupled with research that they have conducted with
intelligent systems and control into the curriculum was ac-
our industrial sponsors, to obtain positions in industry working
complished via a lecture course introducing intelligent control
on intelligent control.
theory, a follow-up lecture-laboratory course, and a parallel
Index Terms—Education, fuzzy control, intelligent control, lab-
senior design project activity. The focus of this paper is on
oratory.
the lecture-laboratory course; however, we do describe several
other details of our curriculum in control.
I. INTRODUCTION
There exist numerous control laboratories around the world
KEY concepts and techniques in the area of intelligent for the instruction of classical, modern, and intelligent control
systems and control were discovered and developed over
techniques; the interested reader is referred to [2] for a
the past few decades [1]. While some of these methods have
thorough exposition and guide to several citations. The novelty
significant benefits to offer, engineers are often reluctant to
of the laboratory course reported herein is in its focus on fuzzy
utilize new intelligent control techniques for several reasons:
control algorithm implementation. In particular, the laboratory
1) there has been a lack of rigorous engineering analysis
demonstrates several key attributes of fuzzy control, including
to verify, for example, stability properties and performance
control design with limited conventional modeling exercises,
characteristics; 2) there is not an established track record for
heuristic construction of nonlinear controllers, comparative
the reliability and robustness of such techniques; 3) there
analysis with conventional controllers, needs/advantages of
has not been enough comparative analysis to determine their
adaptation for fuzzy controllers, and the role of rule-based
advantages/disadvantages relative to conventional methods;
supervisory mechanisms via lectures on complex industrial
and 4) the approaches are not widely understood by practicing
applications. Moreover, we go beyond the treatment of only
engineers. The relative lack of attention given to the potentials
fuzzy control by providing the opportunity to implement neural
of intelligent control, especially in American universities and
networks and genetic algorithms for estimation and control
industry, is cause for some concern, indicating a definite need
(i.e., we include other methods in intelligent control via special
for applications-directed research and education in these areas.
projects for the students based on their interest).
Curricula for control engineering programs has undergone
substantial change in the past 30 years as modern techniques
II. CONTROL CURRICULUM AT THE OHIO STATE UNIVERSITY
for analysis and design find their way into our college courses.
It is quite natural, then, that newer technologies such as intel-
A. Overview
ligent control should be introduced into university curricula.
To set the stage for description of the Intelligent Control
Along with the continuously evolving curricula, there remains
Laboratory course, we briefly describe the Department of
a constant in control engineering education: the recognized
Electrical Engineering’s control systems curriculum at Ohio
need for laboratory experience in the curricula. More and more
State. There is an undergraduate course on control and its
examples of high-quality control laboratories are appearing
corresponding laboratory. At the time of this writing, 16
in universities around the world. Moreover, more and more
graded courses are offered at the beginning graduate (also
educators recognize the importance of a complete educational
available as senior electives) and advanced graduate levels
experience involving theory and practice.
(eight of each). Six of the advanced-graduate level courses
alternate on an every-other-year basis, meaning that 13 courses
run every year. The topics covered in the lecture-only courses
Manuscript received August 30, 1996; revised November 9, 1998. This
work was supported in part by the National Science Foundation under Grant
available for graduate credit are: Filtering and Estimation
EEC 9 315 257.
in Control, Linear Systems, Feedback Control II, Topics in
The authors are with the Control Research Laboratory, Department of
Control Applications (Powertrain Control, or Autonomy in
Electrical Engineering, The Ohio State University, Columbus, OH 43210-
1272 USA.
Vehicles), Nonlinear Systems, Digital Control, Advanced Lin-
Publisher Item Identifier S 0018-9359(99)01244-3.
ear Systems, Stochastic Control, Adaptive Control, Nonlinear
0018–9359/99$10.00 © 1999 IEEE
16
IEEE TRANSACTIONS ON EDUCATION, VOL. 42, NO. 1, FEBRUARY 1999
Control II, Optimal Control, Large-Scale Systems, Robust
applied control research. Moreover, it builds further on an
Control, and Intelligent Control.
existing educational laboratory facility resulting from a NSF
The Control Group at Ohio State is recognized for its
award under the Instrumentation and Laboratory Improvement
instructional control laboratory courses, and has more than
Program (see [5]) for our undergraduate controls laboratory.
ten years experience in their development (see, for example,
The course serves as a complement to the Digital Control
[3]–[5]). Currently, in addition to the undergraduate core
Laboratory course, although neither requires knowledge from
Control Systems Laboratory course, the curriculum offers two
the other. Both courses now run off the same equipment
lecture-laboratory courses at the senior/beginning-graduate
(computers, data-acquisition hardware, and instrumentation).
level. The first is the Digital Control Laboratory course,
Whereas the existing Digital Control Laboratory course em-
a three-credit hour lab/lecture course which emphasizes
phasizes machine-level programming and aspects of hardware,
several important aspects of digital control: instrumentation
the Intelligent Control Laboratory course focuses strictly on
issues (sensors and actuators), microprocessors, operating
aspects of designing and applying intelligent control and
systems, and high-level control languages. The second of
conventional control algorithms to a variety of real processes,
these laboratory courses is the Intelligent Control Laboratory.
with little attention given to details of digital implementation
(principles of A/D and D/A conversion, word length restric-
B. Intelligent Control Sequence
tions, and so on). That is, due to the nature of the experimental
The lecture-only Intelligent Control course mentioned above
testbeds targeted for this project, the majority of the crucial
was introduced as part of a sequence of courses aimed at
control software is already in place, and students are only
bringing current research on intelligent systems and control
required to write control subroutines in C and C
Thus,
into the curriculum. The second course of this sequence is the
many students take both laboratory courses because of their
subject of this paper, and is discussed in the sequel.
differing emphasis.
The two-course sequence (which ran for the first time in
The course consists of two lectures per week, with several
the 1994–1995 academic year) is offered at the graduate level,
hours per week in the laboratory. Important features are:
and is available as a senior elective sequence for motivated
• active participation in current research directions through
undergraduates (where undergraduates complete only a portion
hands-on experimentation with testbeds for comparison
of the laboratory and course). The topical outline for the
of conventional and intelligent control techniques;
three-credit hour lecture-only course is:
• introduction to special purpose hardware (e.g., OMRON
• Fuzzy control: direct, adaptive, and supervisory;
Electronics fuzzy processor) and software (such as OM-
• Neural networks: multilayer perceptron and radial basis
RON software, and fuzzy control toolboxes for Matlab)
function neural network; neural estimation and control;
for intelligent control analysis and design;
• Stability analysis adaptive fuzzy/neural control systems;
• report requirements which will draw from current re-
• Fuzzy/neural systems for identification and estimation;
search, relevant applications, and experimental experience
• Genetic algorithms for computer-aided control design,
in presentation of results.
adaptive control, and estimation.
A portion of the lectures focus exclusively on the labora-
With regard to the treatment of fuzzy control in both
tory experiments (details of coding, research-related issues,
courses, our intent has not been to give an in-depth treatise
modeling, and control objectives), and a portion focus on the
on the theory of fuzzy sets. We have found that electri-
necessary intelligent control design issues. As requirements for
cal engineering undergraduates (through independent research
the course, students are expected to conduct projects (analysis,
projects) have little difficulty in “coming up to speed” in the
design, and application) on several of the testbeds over the
area in a relatively short amount of time. Thus, sometimes
course of the quarter. Required reports not only summarize
as part of this independent study undergraduates take the first
the techniques, procedures and results of the individual exper-
portion of the above course (the direct fuzzy control material),
iments, but also explore additional nuances as directed by the
and then implement fuzzy controllers in the lab to complete
accompanying exercises in the lab notes.
their study.
The graduate students are exposed to the more advanced
B. Syllabus
topics listed above, and at a higher level of sophistication.
Each laboratory section (two sections were conducted in
Several projects in simulation, design, and stability analysis are
Spring 1995) is limited to eight students, and students work
given. Then the graduate students are required to complete the
in pairs. The basic course structure for the ten-week quarter
entire lecture-laboratory course described in the next section.
is as follows.
Finally, we note that other courses in artificial intelligence,
Week 1: Laboratory Software: Exercises for this laboratory
neural networks, and expert systems for monitoring and control
require the student to use Microsoft Windows, Matlab for Win-
are available at OSU.
dows, C programming skills, and Borland C
all of which
III. LECTURE-LABORATORY COURSE
are used in the remainder of the course. Basic procedures
include comparing designs of digital filters in C and Matlab.
A. Course Overview
Week 2: Laboratory Hardware: Students are introduced to
The Intelligent Control Laboratory course capitalizes on
the capabilities of the data acquisition instrumentation. Pro-
the long-standing strength of the Ohio State control group in
cedures include writing routines to access signals from a
YURKOVICH AND PASSINO: LABORATORY COURSE ON FUZZY CONTROL
17
waveform generator, implementing digital filters written in C
5) applications of adaptive fuzzy control: two-link flexible
which interface with real signals, and making comparisons to
robot;
Matlab simulations.
6) applications of adaptive fuzzy control: reconfigurable
Week 3: Matlab Toolboxes: Students
are
introduced
to
control;
special-purpose, commercially available tutorial fuzzy control
7) applications of supervisory fuzzy control: PID autotun-
software for Matlab. Procedures include use of available
ing;
functions and demos, design and simulation of a dc motor
8) applications of supervisory fuzzy control: two-link flex-
fuzzy controller, and various tuning exercises on the developed
ible robot;
controllers.
9) fuzzy versus conventional control: advantages and dis-
Week 4: DC Motor Control: Students gain their first expe-
advantages;
rience of actual implementation. Several dc motor setups with
10) intelligent versus conventional control: advantages and
variable loads are instrumented, offering excellent “starter”
disadvantages.
experiments for students with limited control laboratory ex-
These are weekly lectures of one hour or more in length.
perience. Procedures require students to implement and tune
fuzzy controllers (designed in Week 3) in C code.
IV. LABORATORY TESTBEDS
Week 5: Direct Fuzzy Control: After the first four weeks of
All testbeds are controlled by 486-based PC’s (operating at
introduction, along with the accompanying lectures described
50–66 MHz), introducing a uniformity which is vital in such an
below, students spend the remaining six weeks of the term im-
instructional laboratory. Additional processor hardware (within
plementing controllers on laboratory testbeds (described in the
the PC’s on selected stations) include the OMRON FB-30AT
next section). Students implement a “direct” fuzzy controller
Inference Board, FP-3000 Digital Fuzzy Processor, and FS-
on one of the following: the rotational inverted pendulum,
10AT Inference Software. Data acquisition for each station
ball-beam system, process control plant, inverted pendulum on
is accomplished with the Keithley Instruments DAS-20 card.
an inverted wedge (recently developed experiment), or flexi-
Each PC has an ethernet connection, 16-MB ram, and the
ble arm. Although each testbed emphasizes entirely different
Windows 3.1 operating system.
actuation and sensing technologies, procedures for each are
It is important to note that each testbed emphasizes different
basically the same: design and implement a fuzzy controller
sensing technologies, including optical encoders, potentiome-
(varying from two-input to four-input controllers) and compare
ters, thermal measuring devices, and accelerometers. Actuation
to conventional control designs.
is accomplished with dc and ac motors, including servo-
Week 6: Direct Fuzzy Control: Students repeat procedures
controlled pumps and high-torque motors.
carried out in Week 5, but on a different testbed.
Weeks 7–8: Adaptive Fuzzy Control: Students design and
A. Rotational Inverted Pendulum
implement an adaptive fuzzy control algorithm on one of
the four laboratory testbeds. Each testbed affords interesting
A classic control problem is the inverted pendulum. Most
problems requiring controllers that can adapt to plant
conventional setups consist of a pendulum hinged to a moving
parameter variations so that higher performance control can
cart (driven by a belt or chain) on a linear rail. In this experi-
be achieved. Procedures also include comparison to direct
ment another idea (see [6]) is used in which the pendulum is
(fixed) fuzzy controller designs.
fixed by bearings to a rotating arm.
Weeks 9–10: Project in Intelligent Control: Students carry
This test bed, the result of [7], consists of three primary
out projects which go beyond the first eight weeks of the
components: the plant, digital, and analog interfaces, and the
course in terms of methods applied and control objectives.
digital controller. The overall system is shown in Fig. 1.
A variety of intelligent control (and also estimation or system
The plant is composed of a pendulum and a rotating base
identification) techniques are acceptable, including fuzzy, ex-
made of aluminum rods, two optical encoders as the angular
pert, neural, or genetic algorithms. Application is on one of
position sensors with effective resolutions of 0.2 degrees for
the laboratory testbeds, on the programmable logic controllers
the pendulum and 0.1 degrees for the base, and a large, high-
(complete state-of-the-art PLC units and development systems
torque permanent-magnet dc motor (with rated stall torque of
from Modicon are available), on the OMRON fuzzy processor,
5.15 N-m). As the base rotates the pendulum is free to rotate
or, in special cases, even on a process outside the laboratory.
(high-precision bearings are utilized) through its angle made
with the vertical.
Control objectives for this testbed are twofold: swing-up of
C. Lectures
the pendulum to the vertical position, and then balancing the
There are ten lectures given in the laboratory in parallel
pendulum. Adaptive techniques are required when a weight
with the above labs. These are:
is added to the end of the pendulum, and when additional
1) laboratory orientation, Fuzzy controller C code;
dynamics are added by attaching a half-filled bottle of water
2) fuzzy systems toolbox demonstration;
to the end, which introduces a “sloshing liquid” effect.
3) applications of fuzzy control: surge tank, ball-beam
(video);
B. Ball-Beam System
4) applications of fuzzy control: rotational inverted pendu-
Another classic control problem is to balance a ball in
lum (video);
a groove by tilting a platform on which the ball balances
18
IEEE TRANSACTIONS ON EDUCATION, VOL. 42, NO. 1, FEBRUARY 1999
Fig. 1.
Rotational inverted pendulum.
Fig. 2.
Ball-beam testbed.
(see Fig. 2). Unlike commercially available setups which use
The flexible-link robot testbed consists of a single light-
continuous resistive sensing of the ball position, this apparatus
weight flexible arm counterbalanced with a rigid appendage
was built in-house as an undergraduate research project [8].
(see Fig. 3). The arm is actuated by a dc motor at the
The testbed is also distinguishable from other devices like it in
base, accompanied by its own controller/servo amplifier. A
that it uses discrete sensing of the ball position via a row of 32
high-resolution optical encoder gives angular measurement
phototransistors. Two light sources are used to illuminate the
of the motor shaft, and signal conditioning similar to the
beam from above, so that the ball position is sensed according
pendulum testbed is used prior to sampling. An accelerom-
to its shadow cast on the phototransistors. The angular position
eter mounted at the endpoint of the flexible arm is used to
of the beam is sensed using a potentiometer, while a dc motor
measure linear acceleration at the endpoint. This device is
provides actuation of the beam via a 50 : 1 gear ratio.
produced by Kistler, and has the Kistler Piezotron Coupler
The objective of control experiments is to move the ball
as interface; that output is passed through an analog low-
from one position (at rest) to another position along the
pass filter prior to sampling. A small incandescent bulb is
beam. Adaptation is required when different-sized balls are
mounted on the endpoint, which is used in conjunction with
used on the platform, and also due to inherent nonlinearities
a linear array line-scan camera to record movement of the
(including deadband and backlash introduced by the dc motor
endpoint. The camera system is interfaced with a separate PC,
and gearbox configuration), discrete sensing, and an uneven
and is used solely for displaying endpoint position, not for
rolling surface.
feedback control.
The objective of this testbed is to investigate the ability of
C. Flexible-Link Arm
intelligent control techniques to suppress unwanted vibrations
The single and multiple-link flexible arms serve as excellent
at the endpoint as the arm undergoes large and rapid slews.
testbeds for nonlinear control (in large-angle movements) and
Thus, a typical control experiment is to begin with the arm
vibration suppression control (for endpoint positioning). The
initially at rest, then to slew through an angle of 90 ; feedback
control group at Ohio State has done work with one, two,
variables include the angular position and velocity of the
and three-link flexible robot arms for more than 13 years, and
hub (motor shaft), and the endpoint acceleration. Controller
much of that expertise has been brought into this laboratory
adaptation is required when an extra weight (payload) is
course; see, for example, [9]–[11].
attached to the endpoint.
YURKOVICH AND PASSINO: LABORATORY COURSE ON FUZZY CONTROL
19
Fig. 3.
Flexible-arm testbed.
Fig. 4.
Process control testbed.
D. Process Control Plant
tation problems such as sensor noise, significant time delays,
An often used example for illustration of conventional and
and the lack of a good mathematical model of the plant. The
temperature and level sensors, with heaters and stirrer fans,
intelligent control is the chemical mixing process control
are used in a variety of control objectives on the setup (e.g.,
plant. This testbed consists of four tanks, four liquid-level
temperature or liquid level regulation). Specific problems of
measuring devices, two temperature measuring devices, two
the setup, making accurate control difficult, include accurate
mixers (stirrer fans), and two heaters (see Fig. 4). The liquid
sensing of the liquid level (in the presence of turbulence due
level in each tank is measured by a potentiometer attached
to the pumping action) and deadband nonlinearity in the dc
to a styrofoam float. Temperature measurements are made via
pumps. We can also simulate a pump degradation failure (as
temperature transducers mounted in the “reaction” chamber
if the filters in the pumps get dirty); compensation for this
(where hot and cold liquid is mixed) and in the hot tank,
requires adaptation [12].
which each contain a heater and stirrer. An ac pump is
used to remove liquid from the reaction chamber, while two
dc pumps move liquid from the hot and cold tanks to the
E. Development of New Experiments
reaction tank. The ac pump, heaters, and stirrers may be
We have an on-going effort to develop new experiments that
turned on or off independently, making up five of the plant’s
may be used in the laboratory. Recently, we have completed
inputs. The flow rate of the two dc pumps may be varied
the construction of an inverted pendulum on an inverted
independently by changing their supply voltage (via a PWM
wedge. This experiment presents a significantly challenging
scheme), comprising the remaining two plant inputs.
nonlinear control experiment where we can study control of
The objective of the experiment is to investigate the funda-
an inverted pendulum on an inclined plane (by fixing the
mentals of intelligent process control, with real-life implemen-
inverted wedge position) or balancing of an inverted wedge
20
IEEE TRANSACTIONS ON EDUCATION, VOL. 42, NO. 1, FEBRUARY 1999
(by removing the pendulum). In addition, one can try to solve
these companies has enhanced our research and educational
the simultaneous wedge-pendulum balancing problem. We are
programs (e.g., by providing thesis topics or examples for
also in the process of constructing a magnetically levitated
classroom discussion).
ball that uses a photo-resistive strip as a ball position sensor.
We hold a biannual OSU Control Workshop (which is a
Such experiments are normally constructed by undergraduates
gathering at OSU of roughly 100 professors and graduate
as their senior design projects.
students from Midwest universities) where we provide a “lab-
oratory open house.” These laboratory tours help to facilitate
V. I
the spread of our curricular advancements. Another way that
MPACT OF THE COURSES
we have spread ideas about the courses is through publications
In this section we provide a brief discussion on the impact
at international conferences. See, for example, [14] and [15].
of the course on intelligent control and the intelligent control
laboratory.
VI. CONCLUDING REMARKS
A. Student Reaction
In addressing the need for bringing intelligent control re-
search into the curriculum, we have developed a sequence of
The students generally responded very positively to the
courses accessible to advanced seniors and graduate students
lecture and lecture-laboratory courses. There were over 40
interested in theory and application of intelligent control. The
students that completed the lecture course in Spring 1995
sequence ran with great success for the first time in early
and 16 students who completed the laboratory (with four
1995, with more than 40 students in the lecture course and
undergraduates). There were about 30 students in the lecture
16 students in the laboratory course. It has run regularly
course in Spring 1997 and 12 in the laboratory. The students
since then. Students exiting the sequence are equipped with
provided many positive comments on the student evaluations
the ability to design controllers and estimation/identification
of instruction. For instance, they very much liked the fact that
techniques using fuzzy logic, genetic algorithms, and general
we kept them away from low-level implementation issues and
rule-based systems. Students who also take the laboratory
felt that the course and laboratory were nicely in synch with
course have first-hand experience at implementing real-time
each other. They liked the laboratory lectures and felt that
intelligent control and estimation algorithms, and are well
they helped see how they could apply the methods to even
situated for placement in the job market. It has not been our
more complex industrial applications. Several students made
intention to exhibit documentation and comparative anlaysis
direct use of the intelligent control course and laboratory in
of control algorithms taught and implemented in the course;
their M.S. research and two journal papers were subsequently
the cited references do a complete job of this.
published on some projects done in these classes (students
especially like this). There were a few complaints about
problems with laboratory equipment, but such complaints were
ACKNOWLEDGMENT
minimal as most of the equipment is new. Overall, the students
The authors wish to acknowledge many students who con-
felt that they had been provided a unique and valuable learning
tributed to the development, most notably S. C. Brown and
experience that they would be of benefit to them in their
R. Ordonez. They would also like to thank the Department
careers.
of Electrical Engineering at Ohio State, various other or-
ganizations who discounted or donated equipment for this
B. Textbook, Laboratory Manual
development effort, including OMRON, Modicon, and Battelle
The authors have created a laboratory manual for the
Memorial Institute, and other sponsors or organizations who
intelligent control laboratory that contains all the laboratory
have had the authors give a short course on fuzzy or intelligent
assignments and discusses all the necessary details on how to
control (Air Products and Chemicals Inc., Amoco Research,
complete the laboratories. In addition, the authors have written
Battelle Memorial Inst., Delphi Chassis Div. of General Mo-
a textbook on fuzzy control [13] that includes an instructor’s
tors, Emerson Electric, General Electric Aircraft Engines,
manual (and indicates how to get the laboratory manual that we
General Motors Corp., NASA, General Electric Aircraft En-
use). It is our hope that these publications will help to spread
gines, Reliance Electric, Ford Motor Company, Rockwell
the curricular developments at OSU to other universities.
Int. Science Center, United Technologies Corp., and Wright
Laboratories.).
C. Impact on Industry/Government Laboratories/Universities
REFERENCES
OSU has a long standing tradition of contracted industrial
research in control systems. We have conducted research on
[1] P. J. Antsaklis and K. M. Passino, Eds., An Introduction to Intelligent
and Autonomous Control.
Norwell, MA: Kluwer, 1993.
intelligent control for or with several industries/government
[2] “A world view of control education,” Special Issue of IEEE Control
laboratories or have given short-courses or seminars in in-
Systems, vol. 16, no. 2, 1996.
telligent control for them (please see the acknowledgment
[3] U. Ozguner, “Three-course control laboratory sequence,” IEEE Control
Systems, vol. 8, no. 3, pp. 14–18, 1989.
section for a list). Our graduate students are especially well
[4] S. Yurkovich, D. Clancy, and J. Hurtig, Control Systems Laboratory.
prepared for work with our industrial sponsors due to their
Needham Heights, MA: Simon and Schuster, 1998.
[5] S. Yurkovich, “The instructional control laboratories in electrical engi-
exposure to the theory, application, and implementation of
neering at the Ohio State University,” in Proc. Amer. Contr. Conf., San
intelligent control methods. Moreover, the interactions with
Francisco, CA, 1993.
YURKOVICH AND PASSINO: LABORATORY COURSE ON FUZZY CONTROL
21
[6] K. Furuta, M. Yamakita, S. Kobayashi, and M. Nishimura, “A new
Kevin M. Passino (S’79–M’90–SM’96) received the Ph.D. degree in electri-
inverted pendulum apparatus for education,” in IFAC Advances in Contr.
cal engineering from the University of Notre Dame, IN, in 1989.
Education Conf., 1991, pp. 191–196.
He has worked on control systems research at Magnavox Electronic Systems
[7] S. Yurkovich and M. Widjaja, “Fuzzy controller synthesis for an inverted
Co. and McDonnell Aircraft Co. He spent a year at Notre Dame as a Visiting
pendulum system,” IFAC Contr. Eng. Practice, vol. 4, 1996, to be
Assistant Professor and is currently an Associate Professor in the Department
published.
of Electrical Engineering at The Ohio State University. He is coeditor of the
[8] E. Laukonen and S. Yurkovich, “A ball and beam testbed for fuzzy
book An Introduction to Intelligent and Autonomous Control (Boston, MA:
identification and control,” in Proc. Amer. Contr. Conf., San Francisco,
Kluwer, 1993) and coauthor of the books Fuzzy Control, (Reading, MA:
CA, 1993.
Addison-Wesley, 1998) and Stability Analysis of Discrete-Event Systems (New
[9] S. Yurkovich, F. E. Pacheco, and A. P. Tzes, “On-line frequency domain
York: Wiley, 1998). His research interests include intelligent systems and
information for control of a flexible-link robot with varying payload,”
control, adaptive systems, stability analysis, and fault tolerant control.
IEEE Trans. Automat. Contr., vol. 34, no. 12, pp. 1300–1305, 1989.
He has served as a member of the IEEE Control Systems Society Board
[10] V. G. Moudgal, W. A. Kwong, K. M. Passino, and S. Yurkovich, “Fuzzy
of Governors; has been an Associate Editor for the IEEE TRANSACTIONS ON
learning control for a flexible-link robot,” IEEE Trans. Fuzzy Syst., vol.
AUTOMATIC CONTROL; served as the Guest Editor for the 1993 IEEE Control
3, pp. 199–210, May 1995.
Systems Magazine Special Issue on Intelligent Control and a Guest Editor for
[11] V. G. Moudgal, K. M. Passino, and S. Yurkovich, “Rule-based control
a special track of papers on Intelligent Control for IEEE Expert Magazine
for a flexible-link robot,” IEEE Trans. Contr. Syst. Tech., vol. 2, pp.
in 1996; and was on the Editorial Board of the International Journal for
392–405, Dec. 1994.
Engineering Applications of Artificial Intelligence. He is currently the Chair
[12] J. Zumberge and K. M. Passino, “A case study in intelligent control for
for the IEEE CSS Technical Committee on Intelligent Control and is an
a process control experiment,” in Proc. IEEE Int. Symp. Intell. Contr.,
Associate Editor for IEEE TRANSACTIONS ON FUZZY SYSTEMS. He was a
Dearborn, MI, Sept. 1996, pp. 37–42.
Program Chairman for the 8th IEEE International Symposium on Intelligent
[13] K. M. Passino and S. Yurkovich, Fuzzy Control.
Menlo Park, CA:
Control, 1993 and was the General Chair for the 11th IEEE International
Addison-Wesley, 1998.
Symposium on Intelligent Control.
[14] S. Yurkovich and K. Passino, “An intelligent control laboratory course,”
in Proc. IFAC World Congr., vol. G, San Francisco, CA, pp. 83–88,
July 1996.
[15]
, “A laboratory course for teaching intelligent control,” in Proc.
IEEE Conf. Decision Contr., Kobe, Japan, Dec. 1996.
Stephen Yurkovich (S’79–M’M’82–SM’92) received the B.S. degree in
engineering science from Rockhurst College, Kansas City, MO, in 1978, and
the Ph.D. degree in electrical engineering from the University of Notre Dame,
IN, in 1984.
He is now Professor of Electrical Engineering at The Ohio State University.
His research has focused on the theory and applications of control technology,
in the areas of system identification and parameter set estimation for control,
and fuzzy logic for control, in application areas including flexible mechanical
structures, industrial control systems, and automotive systems. He has been an
author on more than 120 technical publications in journals, edited volumes,
and conference proceedings. He has authored and coauthored the books
Control Systems Laboratory (Kendall/Hunt, 1992), Fuzzy Control (Reading,
MA: Addison-Wesley, 1998), and Control Systems Laboratory (New York:
Simon and Schuster, 1998).
Dr. Yurkovich was Editor-in-chief of IEEE Control Systems from 1993
to 1998. In addition to being General Chair for the 1996 IEEE Conference
on Control Applications, Program Chair for the 1997 American Control
Conference, and General Chair for the 1999 American Control Conference,
he has held numerous positions within the IEEE Control Sysems Society: an
elected member of the Board of Governors from 1994 to 1996, Vice President
for Publication Activities from 1995 to 1996, Vice President for Financial
Activities in 1997, President-Elect in 1998, and President in 1999.
