NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/comp/ai/fuzzy/129

Text File | 1993-01-21 | 46.7 KB | 971 lines

Newsgroups: comp.ai.fuzzy
Path: sparky!uunet!gatech!destroyer!news.iastate.edu!vincent1.iastate.edu!sehari
From: [e-mail] (Farzin Mokhtarian)
Subject: Complete contents of the booklet "Clearly Fuzzy"
Message-ID: <[e-mail]>
Originator: [e-mail]
Sender: [e-mail] (USENET News System)
Organization: Iowa State University of Science and Technology, Ames, Iowa.
Date: Thu, 21 Jan 1993 14:01:34 GMT
Lines: 959


Complete contents of the booklet "Clearly Fuzzy" by:

OMRON Corporation
International Public Relations Section
3-4-10, Toranomon, Minato-ku
Tokyo, 105 Japan

Tel: [phone]
Fax: [phone]

Contact: Tadashi Katsuno

---------------------------------------------------------------------

1. Introduction

Fuzzy Logic is attracting a great deal of attention in the industrial
world and among the general public today. Quick to recognize this
revolutionary control concept, OMRON seriously began to study Fuzzy
theory and technology in 1984, back when the term "Fuzzy" was still
relatively unknown.

Just three years later, OMRON stunned the academic world and triggered
today's boom when it exhibited its first super-high-speed Fuzzy
controller. It was developed jointly with Assistant Professor Takeshi
Yamakawa of Kumamoto University and shown at the Second International
Conference of the International Fuzzy Systems Association (IFSA).

OMRON has since dedicated itself to exploring the potential of this
innovative technology. The company invited Professor Lotfi A. Zadeh,
the founder of Fuzzy theory, to be a senior advisor, and welcomed
researchers from China, a country known as one of the leaders in
Fuzzy Logic study. As a result of technological exchanges with
research institutes of various countries, OMRON's Fuzzy Logic-related
activities are reaching a global scale. Since 1984, OMRON has applied
for a total of 700 patents, making the company an international leader
in Fuzzy Logic technology.

OMRON's enthusiasm for Fuzzy Logic stems from the company's goal of
creating harmony between people and machinery. As a key technology
in OMRON's future, we will be working hard to strengthen and refine
this exciting technology and give it truly useful applications at
production sites, in offices, in public facilities, as well as in
everyday life.

We hope this booklet will be useful in increasing your knowledge,
or at least in sparking your interest in this exciting technology.

                                            OMRON Corporation

------------------------------------------------------------------

2. Truly Friendly Machines

2.1. Arrival of the Fuzzy Boom

The current Fuzzy boom was triggered by the presentation of trial
Fuzzy applications at the Academic Conference of the International
Fuzzy Systems Association (IFSA). The obvious feasibility of these
forerunners of today's Fuzzy Logic deeply impressed conference
attendees. Nowadays in Japan, Fuzzy Logic is successfully being
applied to industrial systems such as elevators and subways and
to an array of consumer electronic products. Convenient Fuzzy Logic
home electrical appliances include washing machines that sense the
dirtiness and type of fabric to automatically determine water flow
and detergent requirements; and vacuum cleaners capable of detecting
not only the presence but the degree of dust on a floor!

2.2. Shades of Gray

The theory of Fuzzy Logic was introduced to the world by Professor
Lotfi A. Zadeh of the University of California at Berkeley.
Professor Zadeh observed that conventional computer logic is
incapable of manipulating data representing subjective or vague
human ideas, such as "an attractive person" or "pretty hot".
Computer logic previously envisioned reality only in such simple
terms, as on or off, yes or no, and black or white. Fuzzy Logic
was designed to allow computers to determine valid distinctions
among data with shades of gray, working similarly in essence to
the processes which occur in human reasoning. Accordingly, Fuzzy
technologies are designed to incorporate Fuzzy theories into
modern control and data processing, to create more user-friendly
systems and products.

2.3. A Warm Welcome in the Orient

Since Fuzzy Logic's world debut 26 years ago, theoretical and
practical studies have been carried out in countries around
the globe; Fuzzy Logic research is currently underway in over
30 nations including the USA, Europe, Japan and China. It may
be surprising to some to note that the world's largest number
of Fuzzy Logic researchers are in China, with over 10,000
scientists and technicians presently hard at work. Japan ranks
second in Fuzzy Logic manpower, followed by Europe and the USA.
Among all nations however, Japan is currently positioned at the
leading edge of Fuzzy Logic application studies. So it may be
that the popularity of Fuzzy Logic in the Orient reflects the
fact that Oriental thinking more easily accepts the concept of
"Fuzziness".

2.4. Fuzzy - Part of Every Day at OMRON

OMRON is also hard at work in the Fuzzy Logic field. Projects
currently on the go at OMRON include working to establish a
Fuzzy technological base, developing new products incorporating
Fuzzy theory, adapting Fuzzy Logic technology to existing
products and conducting seminars for interested audiences
from outside OMRON. Fuzzy Logic has in fact grown to such
proportions that it has become an integral part of the new
corporate culture at OMRON.

-----------------------------------------------------------------------

3. "Fuzzy" Made Clear

3.1. What is "Fuzzy"?

Originally stemming from the fuzz which covers baby chicks, the term
"fuzzy" in English means "indistinct, blurred, not sharply delineated
or focused." This term is "flou" in French and pronounced "aimai" in
Japanese. In the academic and technological worlds, "Fuzzy" is a
technical term. Fuzziness in this sense represents ambiguity or
vagueness based on human intuitions rather than being based on
probability. Twenty six years ago, Professor Lotfi A. Zadeh
introduced "Fuzzy sets" to adapt the concepts of fuzzy boundaries to
science. Fuzzy theory was devised around the Fuzzy sets and a new
field of engineering known as "Fuzzy Engineering" was born. Although
"Fuzzy sets" may sound very mathematical, the basic concept can be
explained simply.

3.2. How Fuzzy Theory Works

o Fuzzy Sets

Let's take an example of the concept "middle age". When we hear the
term "middle age", a certain image comes to mind. But, it is a
concept with fuzzy boundaries which can not be handled by
conventional computers using the binary system. This is where
Fuzzy theory comes in. Let's suppose that we have concluded that
middle age is 45. However, people 35 or 55 years of age can not
be said to be "definitely not middle-aged". There is a feeling,
however, that the implication of "middle age" is somewhat
different inside those boundaries. On the contrary, those younger
than 30 or older than 60 can be considered "definitely not
middle-aged". Such a concept can be represented by a characteristic
function called the "membership function" having a grade between 0
and 1. A Fuzzy set is represented by this membership function.
However, note that the grade within the membership function can be
continuously varied between 0 and 1. This makes possible the
quantitative representation of an abstract intention.

o Crisp Sets

In contrast, the binary system employed in conventional computers
works by first specifying a fixed range, so that "middle age"
represents the age range from 35 to 55 years old. According to
this specification, people who are 34 or 56 years old are not
"middle-aged". Unfortunately, someone who is now considered
young at 34 will suddenly enter middle age as soon as their
next birthday arrives! This sort of unnaturalness is due to
inflexible value assignments. Such concepts with distinct values
of 0 or 1 are called "crisp sets" as opposed to the "Fuzzy sets".

-------------------------------------------------------------------

4. Fuzzy Theory in Action

4.1. Fuzzy Algorithm

One example of Fuzzy theory applications is the handling of
approximate numbers. If approximately 2 is added to approximately 6,
the result will be something around 8. People often make this sort
of calculation. For instance, we frequently estimate the result when
performing a calculation such as "118 + 204." We would say that adding
a number slightly over 100 to another slightly over 200 equals a
number slightly greatly than 300. This sort of calculation comes
easily to human beings but can not be so well handled by conventional
computers, which must have crisp data with which to work.

4.2. The Logic in Fuzzy Logic

Another field that applies Fuzzy theory concerns artificial
intelligence, termed "Fuzzy Logic". One of the differences between
Fuzzy Logic and conventional binary logic is that the truth value
in Fuzzy Logic can be any value between 0 and 1, while that in
binary logic is either 0 or 1. Another difference is that the
Fuzzy proposition includes "fuzziness" as expressed in ordinary
spoken language, in contrast to the crisp proposition which must
be defined distinctly, and is not subject to human intuition.

4.3. Common Sense Fuzzy

"Fuzzy inference" is a reasoning method using Fuzzy theory, whereby
human knowledge is expressed using linguistic rules ("If A is B,
then C is D") with variables B and D. Fuzzy inference is also called
"daily inference" or "common sense inference" since it is performed
by ordinary people. However, conventional computers that employ
binary logic can not handle this reasoning. The use of Fuzzy theory
enables the development of an expert system that can handle
sophisticated knowledge and rich human experience through direct
programming in an almost natural language.

Binary logic based inference is possible only when data coincides
exactly with the premise input. On the other hand, Fuzzy inference
is possible even when the meaning of the fact differs slightly
from the given knowledge. Drawing a conclusion like "Add a little
cold water", Fuzzy inference matches the conclusion based on human
experience, intuition, or possibly even reality.

The "knowledge" part of Fuzzy inference has the structure "if A is
B, then C is D" (example: "If the water is very hot, add plenty of
cold water").  Concepts such as  "very hot"  and  "plenty of cold
water" are subjective and thus represented by Fuzzy sets.

As you may know, Fuzzy theory was devised for the purpose of
enabling machines to handle subjective human ideas and operate
based on advanced knowledge as well as applications of human
beings' intricate experiences. In other words, Fuzzy theory
allows for the development of truly user-friendly machines.

-----------------------------------------------------------------------

5. An Invitation to Fuzzy Control

5.1. The Mechanism: Fuzzy Inference Control

We can examine Fuzzy Control by using the example of controlling an
automobile. In this example, input conditions are speed of the
automobile and its distance to the automobile in front. Amount of
control is expressed in terms of Braking strength.

(1) Express experience and expertise in the form of rules.

With Fuzzy inference control, these rules are called "production
rules". They are represented in the form of "If X is A, then Y is B".
To put it more simply, let's consider two rules as follows:

   Rule 1:  If the distance between two cars is SHORT and the car
            speed is HIGH, then brake HARD for substantial speed
            reduction.
   Rule 2:  If the distance between two cars is MODERATELY LONG and
            the car speed is HIGH, brake MODERATELY HARD (under the
            condition that the front car is moving at a constant
            speed).

(2) Determine membership functions for the antecedent and consequent
    parts.

The distance between the two cars and the car speed (antecedent parts)
and the level of speed reduction, or braking strength (consequent
part), are not numeric values but are represented by "Fuzzy Sets"
expressed through linguistic rules. The distance between the two
cars and the speed have a multiple number of Fuzzy values and are
therefore called "Fuzzy variables". Hence, values (labels) of these
Fuzzy variables and the shapes of membership functions can be
determined. Membership functions (Fuzzy variables) can take three
different shapes: Triangular, Bell-shaped and Trapezoidal.

The shapes differ depending upon the characteristics of the machine
to be controlled. Normally, there are three (large, medium, small),
five (high, moderately high, normal, moderately low, low), or seven
(large, medium and small both in positive and negative directions,
centering around approximately 0) labels. Many Fuzzy controllers
use seven labels, as in the OMRON FZ-3000 Fuzzy Controller, for
example.

(3) Replace linguistic production rules with codes for simpler
    expression.

Although production rules can be expressed with everyday language,
codes are used to simplify the input to the actual Fuzzy Controllers.

(Distance between two cars: X1; speed: X2)
(Braking strength: Y)
(Labels - small, medium, large: S, M, L)

Let's express the above rules using these codes.

    Rule 1:  If X1 = S and X2 = M, then Y = L.
    Rule 2:  If X1 = M and X2 = L, then Y = M.

(4) Execute Fuzzy inference control.

When the rules are programmed into the Fuzzy Controller and it is put
into operation, the Controller will output the most valid control
value based on the variable input conditions.

    1) Establish grades (validity) of input in relation to the Fuzzy
       Sets determined by the rules.

As for the Fuzzy Set (S: short distance) determined by rule 1, the
grade (g11) of the input distance "30m" is 0.4. Similarly, the grade
(g12) of the input speed "40km/H" is 0.2 according to the Fuzzy Set
(M: moderately high speed). As for rule 2, grades (g21) and (g22)
can be determined as 0.7 and 0.6 respectively.

    2) Determine the grade of each antecedent part.

The grade of antecedent parts can be determined by selecting the
smaller value of the grades of inputs. This process is called
"determining MIN (minimum)".

    Rule 1: As g11 = 0.4 and g12 = 0.2, the grade (MIN value) of
            antecedent part (g1) = 0.2.
    Rule 2: As g21 = 0.7 and g22 = 0.6, the grade (MIN value) of
            antecedent part (g2) = 0.6.

    3) Adjust the membership function of the consequent part.

The consequent part of rule 1 is Fuzzy Set (L) representing hard
braking, while that of rule 2 is Fuzzy Set (M) representing medium
(moderately hard) braking. The grades (amplitudes) of these Fuzzy
membership functions are then adjusted to match the grades of their
respective antecedent parts.

    4) Total evaluation of conclusions based on these rules
       (determination of control amounts).

When the conclusions are derived through inference based on each of
these rules (adjusted Fuzzy Sets of the consequent parts), the final
conclusion is then determined by summing the Fuzzy Sets of the
conclusions for each rule. This process is called "determining MAX
(maximum)".

This process considers several variable factors, and is thus very
similar to the human thinking process.

With Fuzzy Control, steps (1) through (4) are performed continuously.
In contrast, with information processing, these procedures are only
executed each time the input data varies.

5.2. The Advantages of Fuzzy Inference Control

o Parallel Control

Conventional control based on modern scientific analysis determines
the control amount in relation to a number of data inputs using a
single set of equations to express the entire control process.
Expressing human experience in the form of a mathematical formula
is very difficult, perhaps impossible. In contrast, Fuzzy inference
control has the following advantages over conventional control:

   1) Expression of control is easy as it need only derive localized
      control rules for each location (or event) in the control range.
   2) It therefore handles complex input/output by using many control
      rules, each of which is effective over a specific location.
   3) Operations can be conducted in parallel (or simultaneously)
      within Fuzzy inference by executing various rules. This
      results in speedy operation, regardless of the total number
      of rules.

o Logical Control

Fuzzy inference control rules are expressed logically using simple
linguistic rules ("If A is B, then C is D"). Because everyday
language can be used, Fuzzy inference control proves ideal for
expressing the sophisticated knowledge of experts and incorporating
valuable intuition (or a "sixth sense").

   1) Multiple conditions can be included as the antecedent part of
      the rules (e.g. If X1 = A, X2 = B and X3 = C, then Y = D).
   2) Rules can be expressed with a single, common format regardless
      of normal or exceptional conditions.

o Linguistic Control

Fuzzy rules can be expressed using everyday language, giving the
following advantages:

   1) Fuzzy control is easy to understand by the machine operator or
      others.
   2) The operator can easily interpret the effect or outcome of each
      rule.

-----------------------------------------------------------------------

6. Growing Up: Fuzzy Technology Catches On

6.1. The Birth and Evolution of Fuzzy

Fuzzy Logic was born only 26 years ago when Professor Lotfi A. Zadeh
submitted a paper entitled "Fuzzy Sets" to the science magazine
"Information and Control". In that paper, he labeled sets with unclear
boundaries "Fuzzy sets," such as attractive people, tall people, and
large numbers. According to Dr. Zadeh, the Fuzzy set plays an important
role in pattern recognition, interpretation of meaning, and especially
abstraction, the essence of the thinking process of the human being.

6.2. Is "Fuzziness" Really Better?

Dr. Zadeh was one of the original founders of the modern control theory
and remains an authority in this field. Modern control theory is exact,
precise, and logical, harboring no hint of "fuziness".

Today, however, the subjects of control have become increasingly larger
in scale, in turn requiring more advanced and complex control systems,
like those used to control robots and rockets. You need a tremendous
amount of power if you want to use a computer to execute such
complicated control using modern theory. Precise programming is
needed for every instruction and every piece of data to put the
computer into operation. It also takes an extremely long time to
execute the programs. Dr. Zadeh devised Fuzzy theory to overcome
these debilitating limitations of modern theory.

There was also another, probably more important factor that encouraged
him to come up with a new idea. Conventional computers work by
identifying the factor which seems to have the strongest influence on
the systems to be controlled, since it is impossible to simultaneously
command all the factors that affect the system. In other words, the
computer assumes that the system only consists of those selected items.
Moreover, all assumed factors must be described digitally. So for some
items which are unclear, the computer simply assigns an appropriate
value. The computer is, of course, capable of accurate and fast
computation. However, as the conditional parameters include many
hypotheses, the computer may sometimes yield a ridiculous conclusion
contrary to what common sense would lead us to expect. This is caused
by its attempts to replace "fuzziness" with fixed numeric values.
Thus, it became necessary to develop a theory capable of dealing
with the vagueness prevalent in everyday decisions.

6.3. Strong Opposition

Even though Dr. Zadeh's theory is now quite popular and quoted in a
large number of academic papers, it had to endure skepticism and
hostility from US researchers and academics in its early days.
Some American mathematicians scoffed at the theory, saying that
"fuzziness" could be represented using conventional mathematics.
Once a noted authority in modern theory, Dr. Zadeh's ready
acceptance of "fuzziness" was considered to be a frivolous
escape from his own beliefs, and many criticized him for not
fulfilling his duty as a scientist.

6.4. A Profile of Professor Zadeh

You may want to know a little about the Professor. Here is a
very brief profile:

Lotfi A. Zadeh was born in Iran on February 4, 1921. In 1956,
he was a visiting member of the Institute for Advanced Study in
Princeton, New Jersey and held numerous distinguished visiting
appointments around the US. In 1959 he joined the University of
California's Electrical Engineering Department at Berkeley, and
served as its chairman from 1963 to 1968.

Before 1965, Dr. Zadeh's work focused on system theory and
decision analysis. Since then his interests have shifted to
the theory of Fuzzy sets, and its applications.

Zadeh attended the University of Teheran, MIT, and Columbia
University, and is a fellow of the IEEE and AAAS. He is also
a member of the National Academy of Engineering. Now, Dr.
Zadeh is a senior advisor to OMRON Corporation.

6.5. A Motivating Debate

Here is a little story about how Fuzzy Logic was invented. One
day, Dr. Zadeh got into a long argument with a friend about who
was more beautiful, his wife or his friend's. Each thought his
own wife was more beautiful than the other's wife. There is,
of course, no objective way to measure beauty. The concept of
"beautiful" greatly differs among people. Although they continued
the argument for a long time, they could not arrive at a
satisfactory conclusion. This argument triggered Dr. Zadeh's
desire to express concepts with such fuzzy boundaries
numerically, and he thereby devised Fuzzy sets. Thus goes the
legend.

6.6. From Industry to Consumer

The first applications of Fuzzy theory were primarily industrial,
such as process control for cement kilns. Then, in 1987, the
first Fuzzy Logic-controlled subway was opened in Sendai in
northern Japan. There, Fuzzy Logic controllers make subway
journeys more comfortable with smooth braking and acceleration.
In fact, all the driver has to do is push the start button!
Fuzzy Logic was also put to work in elevators to reduce
waiting time. Since then, the applications of Fuzzy Logic
technology have virtually exploded, affecting things we use
every day.

Major Areas of Fuzzy Research and Applications

Field              Major Applications

Automation         Steel/iron manufacturing, water purification,
                   manufacturing lines and robots, train/elevator
                   operation control, consumer products, etc.

Instrumentation    Sensors, measuring instruments, voice/character
and analysis       recognition, etc.

Design/judgement   Investment/development consultation, train
                   scheduling, system development tools,
                   trouble-shooting, etc.

Computers          Operators, arithmetic units, microcomputers,
                   industrial calibrators, etc.

Information        Database, information retrieval, system
processing         modelling and mathematical programming, etc.

6.7. Historically Speaking ...

The year 1990 witnessed the 25th anniversary of the invention of
Fuzzy theory. It has undergone numerous transformations since its
inception with a variety of Fuzzy Logic applications emerging in
many industrial areas. Dividing these past years into different
stages, the early 1970s are the "theoretical study" stage, the
period from the late 1970s to early 1980s the stage of "developing
applications for control", and that from late 1980s to the
present the stage of "expanding practical applications".

Here are the major events in the history of Fuzzy Logic:

1965:  Professor L. A. Zadeh of the University of California at
       Berkeley introduces "Fuzzy sets" theory.
1968:  Zadeh presents "Fuzzy algorithm".
1972:  Japan Fuzzy Systems Research Foundation founded (later
       becoming the Japan Office of the International Fuzzy
       Systems Association (IFSA)).
1973:  Zadeh introduces a methodology for describing systems
       using language that incorporates fuzziness.
1974:  Dr. Mamdani of the University of London, UK succeeds
       with an experimental Fuzzy control for a steam engine.
1980:  F. L. Smidth & Co. A/S, Denmark, implements Fuzzy
       theory in cement kiln control (the world's first
       practical implementation of Fuzzy theory).
1983:  Fuji Electric Co., Ltd. implements Fuzzy theory in the
       control of chemical injection for water purification
       plants (Japan's first).

1984:  International Fuzzy Systems Association (IFSA) founded.
1985:  1st IFSA International Conference.
1987:  2nd IFSA International Conference. (Exhibit of OMRON's
       Fuzzy controller, a joint development with Assistant
       Professor Yamakawa).
       Fuzzy Logic-controlled subway system starts operation
       in Sendai, Japan.
1988:  International Workshop on applications of Fuzzy Logic-
       based systems (with eight Fuzzy models on display).
1989:  The Laboratory for International Fuzzy Engineering
       Research (LIFE) established as a joint affair between
       the Japanese Government, academic institutes and
       private concerns.

       Japan Society for Fuzzy Theory and Systems founded.

------------------------------------------------------------------

7. A Fuzzy Future

7.1. Fuzzy Fever Hits Japan

1987 marked the start of Japan's so-called "Fuzzy boom", reaching
a peak in 1990. A wide variety of new consumer products since then
have included the word "Fuzzy" on their labels and have been
advertised as offering the ultimate in convenience.

For instance, Fuzzy Logic found its way into the electronic fuel
injection controls and automatic cruise control systems of cars,
making complex controls more efficient and easier to use. The
"Fuzzy" washing machine has more than 400 preprogrammed cycles;
yet despite this technological intricacy, operation is very
simple. The user only needs to press the start button and the
rest is taken care of by the machine. It automatically judges
the material, the volume and the dirtiness of the laundry and
chooses the optimum cycle and water flow. In air conditioners,
Fuzzy Logic saves energy because it starts cooling more
strongly only when a sensor detects people in the room.

We could go on and on with examples of camcorders, television
sets, and even fund management systems. The sweeping
popularity of Fuzzy Logic in Japan might even surprise
Dr. Zadeh, its founder.

7.2. No Limits: Promise for the Future

Just from these few examples, it is clear that Fuzzy Logic
encompasses an amazing array of applications. Fuzzy Logic can
appear almost anyplace where computers and modern control
theory are overly precise; as well as in tasks requiring
delicate human intuition and experience-based knowledge.
Now that your mind is open to Fuzzy thinking, here are some
unique ideas applying Fuzzy Logic.

7.3. "Fuzzy" Child Care Expert System

Here is an idea a 24-year-old housewife developed from her
experience in raising children. It may seem obvious that
babies don't drink the way it is described in child care
books. They may drink a little or a lot depending on their
physical condition, mood, and other factors. She conceived
a Fuzzy Logic program that would recommend how much to feed
the baby. The program determines the appropriate amount of
milk according to a knowledge base that includes the child's
personality, physical condition, and some environmental
factors. Although adapting Fuzzy Logic to babies may seem
silly, one can easily imagine using it to control the
feeding of animals in captivity, for instance.

7.4. Fuzzy is for Everyone

Many ideas have been derived from everyday activities in the
home, like the Fuzzy ventilation system. It uses Fuzzy Logic
to switch a fan on and off as dictated by its knowledge base
of the amount of smoke, odors, and room temperature and
humidity. The Fuzzy bath, for example, has a controller that
keeps the temperature of the water just right, not too hot
and not too cold. If the water is lukewarm at first, it adds
heat at a slower rate than if it's cold, avoiding wasteful
overheating.

With the right Fuzzy outlook, you could be the next to
discover another innovative application of Fuzzy Logic.

------------------------------------------------------------

8. OMRON and Fuzzy Logic

OMRON is renowned worldwide for its leading-edge Fuzzy Logic
technology research and applications. What has this
technologically advanced company achieved and how? What does
the future hold for this exciting Fuzzy Logic? Through an
interview conducted in February 1991 with General Manager
Masayuki Oyagi of OMRON's Fuzzy Technology Business Promotion
Center, we hope to answer these questions.

Q. How did OMRON become involved with Fuzzy Logic technology?

A. In the early 1980s, we were fortunate enough to meet
   Assistant Professor Takeshi Yamakawa of Kumamoto University
   who specialized in this peculiar new technology known as
   "Fuzzy Logic". Our difficulties in control applications with
   conventional solutions, combined with his enthusiasm for
   Fuzzy Logic's abilities, led us to start studying it, but
   with only a few researchers. The late Executive Advisor
   Kazuma Tateisi (then Chairman), however, was most impressed
   with Fuzzy Logic and correctly predicted its importance.
   His encouragement led to the formation of the Fuzzy Project
   team, now the Fuzzy Technology Business Promotion Center,
   which conducts basic studies and explores new business
   opportunities.

Q. OMRON's R&D efforts have given rise to numerous original
   applications for Fuzzy Logic. Could you give some examples?

A. The most obvious example would be the Fuzzy controller, the
   first of its kind in the world. Developed in conjunction
   with Professor Yamakawa, this breakthrough was a huge
   sensation at every academic conference and fair it was
   exhibited at. Several varieties of Fuzzy controllers are
   already on sale on the Japanese market. There are also
   Fuzzy temperature controllers and Fuzzy software
   development support tools to assist programmers.

   To give some interesting applications, we developed a robot
   which can grasp something "pretty" soft and fragile - tofu
   (bean curd); and a can sorting machine capable of
   identifying cans by color. Overall, OMRON has more than
   100 successful applications, 20 of which are now available
   to the public.

   As 1991 progresses, you can expect more OMRON Fuzzy Logic-based
   products to be introduced. To date we have applied for more
   than 700 patents, a figure that gives some indication of
   OMRON's strength in Fuzzy Logic applications.

Q. Fuzzy Logic technology is obviously important to OMRON. What
   degree of importance does it have within the company?

A. In President Yoshio Tateisi's 1991 New Year address to OMRON
   employees, Fuzzy Logic was identified as one of our core
   technologies for the 1990s. By 1994, over 20% of our entire
   product line will include some form of Fuzzy Logic.
   Considering the diversity of OMRON's products, this is a
   challenging and significant goal.

   OMRON's R&D investments account for approximately 7% of its
   total sales and I think Fuzzy Logic research represents
   nearly 1%.

Q. OMRON is not alone in the Fuzzy Logic business. How does it
   distinguish itself from its competitors?

A. One of the main characteristics of OMRON's Fuzzy Logic-related
   business is the completeness of its product line. OMRON is
   presently the only company which provides an entire range of
   Fuzzy Logic products, including digital and analog units, at
   virtually every speed, inference scale and computation capacity.
   OMRON also offers Fuzzy Logic products in complete sets,
   including chips, software, and development tools, which can be
   used both in-house and by customers. Almost eight years of
   experience with Fuzzy Logic have gone into all of these products.

   There are an amazing number of beneficial Fuzzy Logic applications
   bearing OMRON's name, both original and joint customer
   development projects; the largest number in the world, I think.
   This success lets us continue to satisfy each customer's
   particular needs.

Q. Aside from being a fascinating technology, what makes it so
   attractive?

A. OMRON doesn't think Fuzzy Logic itself makes products better.
   What is more important is the quality of user benefits that
   Fuzzy Logic can offer. Any business operates towards goals,
   such as major performance improvements, cost reductions,
   miniturizing, or others. To attain these goals, businesses
   will usually refine their operations, generally without
   concern for the kind of technology used. But they do care
   about whether the technology can really work for them. Where
   existing computers function perfectly, such as for wage
   calculation, Fuzzy Logic has no value. However, with
   applications that are difficult or impossible  using
   conventional technology, Fuzzy Logic may be the answer.

Q. Where does Fuzzy Logic exhibit an improvement over previous
   technology?

A. The basic characteristic of Fuzzy Logic is that it can handle
   information with unclear boundaries, at any stage of input,
   processing, computation, memory or output. In other words, it
   can manage "fuzziness". The logic itself is purely mathematical,
   so the results are not "fuzzy" but rather very clear and precise.

   Consider the can sorting machine which I mentioned earlier. With
   Fuzzy Logic, a computer can be instructed to sort cans according
   to their colors such as "reddish" or "bluish", instead of by
   reading characters printed on labels. Certainly character
   recognition technology for reading labels is very advanced, but
   when the can is turned so that the label isn't visible, it can't
   work. This is exactly where Fuzzy Logic is best.

Q. What else is happening with Fuzzy Logic at OMRON? How many people
   are involved with this technology?

A. I'm not sure of the exact number but research on the technology
   itself in addition to developing applications involves many
   people. As an indication, at least 1,000 people have taken a
   Fuzzy Logic seminar.

   Some are members of the Laboratory for International Fuzzy
   Engineering Research (LIFE). One person from our Fuzzy Technology
   Business Promotion Center is now working at OMRON Advanced
   Systems, Inc. in Silicon Valley, studying American technology as
   well as introducing Japanese technology to the US staff. We are
   also planning joint studies with various overseas manufacturers
   and seminars are held regularly, probably weekly, for both OMRON
   employees and our customers. Although most of these activities
   are within Japan, we plan to expand them to other countries this
   year.

   The first product scheduled for marketing abroad in 1991 is the
   Fuzzy temperature controller, to be introduced at the upcoming
   Hanover Fair. This will be followed by the Fuzzy chip. OMRON will
   continue its marketing efforts overseas with Fuzzy Logic products,
   ultimately aiming for simultaneous worldwide release. This coming
   spring, a Fuzzy Logic product showroom will open at OMRON
   Electronics, Inc. in Schaumburg, Illinois.

Q. That explains OMRON's aggressive marketing strategies. Some people,
   however, say that  Fuzzy Logic in the  US and  Europe is not as
   popular as in Japan, partially due to the term "Fuzzy". What is
   your impression?

A. I think there are positive and negative feelings about this term.
   In its early days, "Fuzzy" was not considered an academic term.
   Because of this, however, people got the impression that this
   technology was something quite singular which, I think, gave it
   more impact. On the down side, people thought that its results
   or ability would be "fuzzy", and questioned the product
   reliability.

   Regardless of that, the fact is that Fuzzy Logic is used in very
   demanding areas, including nuclear power plants. In the US, NASA
   is working to implement Fuzzy Logic control in space environments,
   an exceptionally difficult task. There are many energetic Fuzzy
   Logic researchers in the US, Europe, and other places, which is a
   favourable change from earlier criticism of this unique
   technology. In fact, American trade magazines are constantly
   asking us for interviews, and French and German groups have been
   visiting OMRON regularly since 1989. This makes me confident that
   Fuzzy Logic technology will grow rapidly in both US and Europe in
   the near future.

   If consumer electronics giants such as GE introduce products with
   Fuzzy Logic, you may see a boom even larger than the one
   experienced in Japan last year. New technology that can handle
   things conventional machines can not, will naturally surprise
   and excite people, in any market and in any country.

Q. 1990 in Japan was considered the "year of Fuzzy Logic". What was
   OMRON's part in that and what are your comments on the boom?

A. With Fuzzy Logic, OMRON's goal is to raise the functions and
   capabilities of machines to levels comparable to human beings.
   In a sense, it can be considered "Artificial Intelligence" (AI).
   The left hemisphere of a human brain is used for logical
   processes, like reading and talking, while the right hemisphere
   is for intuitive and emotional mechanisms as well as unconscious
   information processing. Conventional computers imitate the left
   side, while Fuzzy Logic plays the role of the right side.

   In chess, for instance, players make instant judgments, which
   would take many hours with a conventional computer. Such
   advanced thinking is the product of the cooperative efforts of
   both sides of the brain. We are imitating real life and are
   working on integrating conventional computers with Fuzzy Logic,
   expert systems, neural networks, and other technologies.
   OMRON's goal is to create machines that approximate human
   intelligence and capabilities, and yet still be compact and
   inexpensive.

   The 1990 Fuzzy Logic boom, I think, was the first wave which
   accurately reflected the direction of the technology and it
   motivated us to go further. The market's enthusiastic response
   was due to its sense that this long-awaited technology would
   create truly intelligent, user-friendly machines.

-----------------------------------------------------------------

9. Main Events at OMRON Related to Fuzzy Logic Technology

1984  Research into Fuzzy Logic begun.
1986  Fuzzy Logic medical diagnosis system introduced.
1987  Assistant Professor Takeshi Yamakawa of Kumamoto University
      (now Professor of Kyushu Institute of Technology) introduces
      super high-speed Fuzzy controller, test-manufactured by
      OMRON, at the 2nd Conference of the International Fuzzy
      Systems Association.
1988  World's first super high-speed Fuzzy controller, FZ-1000,
      marketed.
      OMRON participates in the establishment of Laboratory for
      International Fuzzy Engineering Research.
      F (Fuzzy Logic technology research and marketing) project
      formed.
      OMRON participates in the Fuzzy Logic research project of
      the science and Technology Agency.
      Four working models of Fuzzy Logic systems displayed at
      the international workshop on Fuzzy Logic applications.
1989  Professor L. A. Zadeh welcomed to OMRON as senior advisor.
      Ten new products using Fuzzy Logic technology introduced,
      including chips, controllers, and software.
      Fuzzy Technology Business Promotion Center established.
      Bank note feeding mechanism using Fuzzy Logic developed
      for ATMs.
      Fuzzy hybrid control method developed.
1990  "LUNA-FuzzyRON" Fuzzy Logic software development support
      system developed.
      Fuzzy Logic human body sensor developed.
      Fuzzy controller related gain adjustment method devised.
      Failure diagnosis and prediction system for machine tools
      developed using Fuzzy Logic expert system.
      Fuzzy inference unit, C500-FZ001, marketed.
      Two new series of digital Fuzzy processors developed,
      FP-3000 series controllers and FP-5000 series multitask
      processors.
      Development tools for the FP-3000 marketed.
      Fuzzy inference molding machine support system developed.
      Fuzzy temperature controller, E5AF, marketed.

---------------------------------------------------------------

10. Fuzzy Logic Products

OMRON has released numerous innovative products that use Fuzzy
Logic. A few of those products scheduled for release overseas
are listed below:

o FP-3000 Digital Fuzzy Processor-Controller

Cost-effective Fuzzy chip ideal for control and simple pattern
recognition.

* High-speed inference processing - 650 u-s/(5 antecedents and
  2 consequents, 20 rules, 24 MHz (external clock speed)).
* Bus interface similar to that of an SRAM allows connection to
  various CPUs.
* Fuzzy Logic operation can be accomplished on a single chip
  (Single mode).
* High 12-bit resolution.
* Up to 128 rules applicable for each inference (Expanded mode).

o FS-10AT Fuzzy Software Tool

A personal computer software designed to create rules and
membership functions for Fuzzy inference.

* Compatible with IBM PC-AT.
* Allows performance of trial control using A/D and D/A
  expansion boards.
* Outputs created rules and membership functions as object
  code for the FP-3000 Fuzzy controller and FB-30AT Fuzzy
  inference board.

o FB-30AT Fuzzy Inference Board

FP-3000 chip-packaged board

* Can be inserted into an IBM PC-AT expansion slot.
* Uses the rules and membership functions created by the FS-10AT.
* Provided with driver software, allows Fuzzy inference to run
  with the user's software.
* Applications include evaluation and field tests of the FP-3000,
  and addition of Fuzzy Logic functions to personal computers.

o E5AF Fuzzy Temperature Controller

The industry's first temperature controller to employ Fuzzy Logic.

* Highly precise (+/- 0.3% error) and fast response to external
  disturbance.
* Hybrid control integrates PID control and Fuzzy Logic control to
  improve disturbance response significantly (50% higher than
  conventional PID control).
* Easy operation - similar to that of conventional models.
* Automatic Fuzzy Logic parameter setting. Fuzzy Logic parameters
  can be programmed to fit the application.
* Ideal for use in physical/chemical equipment, industrial
  furnaces, and semiconductor manufacturing equipment.

------------------------------------------------------------------

11. Fuzzy Logic Technologies

OMRON is also involved in regular development of practical Fuzzy
Logic applications. Here are some examples:

o Fuzzy Logic Failure Diagnosis and Prediction System for Machine
  Tools

In a joint development with Komatsu Ltd., this system generates
and displays various machine failure predictions in order of
probability, enabling a much simpler detection of the real cause
of the fault. It will reduce servicing time by 24%, and software
development time to 1/5 of conventional systems.

o Fuzzy Inference Molding Machine Support System

This system uses Fuzzy inference to automatically remedy the
conditions that cause plastic molding failures. Unlike
conventional systems which call for expert attention, this new
system only needs a simple defects input into the built-in
controller. Fuzzy inference, with its expert knowledge base,
takes care of the rest automatically, and at the same level
of competence as a specialist.

o Bank Note Feeding System Employing Fuzzy Logic for ATMs and CDs

The texture and quality of bank notes stored in automatic teller
machines (ATMs) and cash dispensers (CDs) are easily affected by
ambient humidity, conveyance conditions, etc., which in turn makes
stable bank note feeding difficult. With the aid of Fuzzy Logic,
this new mechanism keeps the gap between the rollers at the
optimum level, notably increasing the reliability of ATMs and CDs
as well as reducing the need for maintenance.

--