Opening Ceremony
Date: October 24 (Tuesday)
Place:224 Room in Nagoya Conference Center
Time Table:
    13:00-13:30    Opening Ceremony
    13:40-15:55    Plenary Speech
Dr.  Toshiharu Aoki (13:40-14:25)
Prof.  Guenther Schmidt (14:25-15:10)
Mr. Jerome Clayton Glenn (15:10-15:55)
(15 minutes Break)
    16:10-17:30    Industrial Forum (Panel discussion)
Topics:    What is Next Target of Industrial Technologies in the 21st Century ?
Chair:    Fumio Harashima
Speaker:     T. Aoki, G. Schmidt, J. C. Glenn, R. C. Luo and K. Furuta (5 persons)
Contents: 5 minutes presentation per person and panel discussion after these presentation

Plenary Speech

Dr. Toshiharu Aoki
President & CEO
NTT DATA CORPORATION

Title: Changes in Industrial Structure and Intellectual Creation Activity Brought about by the Expansion of the Internet

Abstract

The expansion of the Internet is about to result in dramatic changes in society.  These changes will produce three ideal situations :
(1) Creation of a "perfect market," where suppliers and consumers are on an equal footing.
(2) Creation of a "perfect community," free from space or time constraint.
(3) Creation of a "perfect value," respecting the sense of value of individuals.

 In particular, these three ideal situations will transform the very foundation of commercial transactions into one which encourages the emergence of previously undreamed-of businesses, leading to formidable changes that will engulf the whole industrial structure. These changes will also have far-reaching implications for "research"  and  "technical development," or ,in short, the activity of "intellectual creation"  which we have practiced.  They will completely redefine the process of such activity and the meaning of creativity.

Anyone will be able to share in intellect and knowledge equally, which have traditionally been unevenly distributed, and to interact with any intellectual communities. Such a new environment will provide the opportunity for a completely different intellectual activity.

The recent move to claim intellectual property rights for business models symbolizes the emergence of a new perception of the value placed on creativity.

I will speak of the "change in industrial structure" and the "change in intellectual creation activity"  that will be brought about by the expansion of the Internet.
 

Dr. Guenther Schmidt
Professor
Institute of Automatic Control Engineering (LSR)
Technische Universitaet Muenchen

Title: Industrial Automation at the Brink of the 21st Century

Abstract

Global markets,  changing customer demands and progress in process,  manufacturing and information technologies will bring about new ways of looking at current automation approaches and technologies. The presentation will discuss how improvements in control and automation systems'  preformance and value can be achieved by borrowing and integrating technologies from areas such as computers, telecommunications, or even e-commerce and e-business.  This development will also have major impact on management of automation projects and education of automation engineers.  As a result a mustation of classical Industrial Automation into the broader field of Industrial Information Technology and Automation can be expected in the near future.
 
 

Dr. Jerome Clayton Glenn
Executive Director
The American Council for the United Nations University

Title: Strategic Directions for Science and Technology in the 21st Century

Abstract

The Millennium Project of the American Council for the United Nations University has created da network of 600 futurists, scholars, business planners, scientists, and policy makers who work for the United Nations and other international organizations, governments, corporations, NGOs, and universities in 50 countries who provide a futuristic and contemporary assessment of the global challenges we face. This speech will highlight the implications of their research on 15 global challenges for science and technology strategic directions for the 21st Century from the "State of the Future at the Millennium"  co-authored by the speaker.

The speech will also discuss Millennium Project's work in progress on future international issues in science and technology for the U.S.  Department of Energy as well as science and technology questions from very long-range scenarios for the Foundation For the Future. In conclusion,  the post-information age synergies of artificial intelligence, genetic engineering, nanotechnology, communications technology, biotechnology forming the "Conscious-Technology"  age will be discussed drawing from "Future Mind" ("Future Heart"-TBS Japanese 1993 edition) by speaker.
 

Tutorial

Tutorial Schedule

October 23 (Monday)
13:30-16:00
T1) Neural-Fuzzy System for Motor Incipient Fault Detection and Diagnosis
    Prof. Mo-Yuen Chow,　Room 222
T2) Information Aggregation in Intelligent Systems　
    Prof.  Imre J. Rudas　Room 223
T3) Nano・Technology I
    Prof, W, R, Fahrner and Prof.  U. Hilleringmann, Room 231

October 24 (Tuesday)
9:30-12:00
T4) Sensorless Control of Induction Motors
    Porf. Joachim Holtz, Room 222
T5) Application of Computational Intelligence in Mechatronics
    Prof. Okyay Kaynak, Room 223
T6) Nano-technology II
    Prof.  U. Hilleringmann and Prof. W. R. Fahrner,　Room 231

October 25 (Wednesday)
9:30-12:00
T7) The Internet, Electronic Commerce, and Factory Automation
    Prof, Alfred C. Weaver, Room 134

October 26 (Thursday)
T8) Soft Computing and its Application
    Prof. B. M. ‘Dan’ Wilamowski, Room 223

October 27 (Friday)
T9) Bond Graph and Related Topics
    Prof. Jean Thoma, Prof. Kazuhiro Tanaka and Dr. Katsuya Suzuki, Room 233

Oct. 23 (MON) 13:30-16:00
T1) Neural-Fuzzy System for Motor Incipient Fault Detection and Diagnosis
    Prof. Mo-Yuen Chow, North Carolina State University

Induction motors are the workhorse of our industry because of their versatility and robustness.  The use of induction motors in today's industry is extensive and they can be exposed to different hostile environments, misoperations, and manufacturing defects.  Thus motor fault detection and diagnosis is one of the major concerns in today’s industry.

This tutorial will provide:
Overview of motor failure, reliability, and availability concepts;
Overview of several popular motor incipient faults and symptoms;
Overview of Set Theoretic and its application to organize information in motor fault detection and diagnosis;
Overview of Feedforward Artificial Neural Networks;
Overview of Fuzzy Logic memberships functions and fuzzy rules;

Discussion of a Neural-Fuzzy system, of which the process can be easily interpreted heuristically after successful neural network training, with its application to motor incipient fault detection and diagnosis.

This tutorial will use induction motor insulation failures and bearing faults as examples to illustrate the use of Set Theoretic, Neural Networks and Fuzzy Logic technologies for motor incipient fault detection and diagnosis.

Oct. 23 (MON) 13:30-16:00
T2) Information Aggregation in Intelligent Systems
Prof.  Imre J. Rudas, Budapest Polytechnic

Information aggregation is one of the key issues in the development of intelligent systems, like neural networks, neuro-fuzzy systems, fuzzy knowledge based systems, vision and decision making systems, etc.  Fuzzy set theory provides a host of attractive operators for integrating the membership values representing uncertain information.  The variety of these operators might be confusing and make it difficult to decide which one to use in a specific model or situation.

The aims of the tutorial are:
　to give a survey of the existing aggregation connectives starting from the classical Zadehian-operators, through the theory of t-operators, compensative and averaging operators till the most up-to-date operators, containing the results of the author and his colleagues on entropy based and evolutionary operators,
　to give the most important criteria for selecting appropriate aggregation operators.
　performance evaluation of operator families in control systems by simulation.

Oct. 23 (MON) 13:30-16:00
T3) Nano-Technology I
Prof.  W. R. Fahrner, University of Hagen
Prof.  U.  Hilleringmann, University of Paderborn

Some quantum mechanical laws governing the nanoworld are introduced.

The process sequences for production of nanomaterials are demonstrated.  Chemical molecular synthesis is explained.  The results are compared to computer simulations.  It is shown how wheel-like structures can be produced and combined to gears and transmissions.  Similarly,  tube-like and channel like structures are produced in the simulations.  It is explained how nanodefects are formed in semiconducting material. Finally it is discussed which tools are available to analyze structures in the nanometer range.  X-ray diffraction, decoration, Raman, and spectral ellipsometry as characterization tools of nanostructures are presented.  As exemplary materials, carbon (nanotubes), nanocrystalline silicon, and zeolites are discussed.

Typical nanoprocesses to obtain and to structure nanodevices are presented.
(i) oxidation, ion implantation, physical vapor deposition, (reactive) sputtering, (plasma enhanced) chemical vapor deposition,  and the various forms of epitaxy such as GPE,  LPE, and MBE
(ii) lithographic processes
(iii) reactive ion etching
(iv) focussed ion
(v) implantation masks made of crystals with extremely large channel diameters,
(vi) the techniques for internal gettering
(vii) The smart cut (exfoliation) as obtained by hydrogen implantation for SOI structures
(viii) the LIGA technique.

Oct. 24 (TUE) 9:30-12:00
T4) Sensorless Control of Induction Motors
Prof. Joachim Holtz, Wuppertal University

An insightful interpretation of the dynamic behavior of induction machines is given using the complex notation. The approach enables the visualization of ac machine dynamics through complex signal flow graphs. These simple structures assist to form an understanding of the internal dynamic processes of a machine and their interactions with external controls. The approach is subsequently used for an introduction to sensorless ac drive control.

Controlled ac drives without mechanical sensors for speed or motor shaft position have the attraction of lower cost and higher reliability. To perform sensorless control, the estimation of internal state variables of the machine is required. The assessment is exclusively based on measured terminal voltages and currents. A variety of sensorless controlled ac drive schemes have been developed for practical application. Low cost, medium performance sensorless drives can be designed using simple algebraic speed estimators. High-performance systems rely on dynamic models for the estimation of the magnitude and spatial orientation of magnetic flux waves in the stator or in the rotor. It is shown that the accuracy of estimation is much improved if the nonlinearities of the inverter are adequately modeled. Open loop estimators and closed loop observers differ with respect to parameter sensitivity, robustness, and limits of applicability. The robustness of a sensorless ac drive can be improved by adequate control structures and by parameter identification techniques.

Oct. 24 (TUE) 9:30-12:00
T5) Application of Computational Intelligence in Mechatronics
Prof. Okyay Kaynak, Bogazici University

Intelligence in the form of well-organized solutions to the ill-posed problems has been the primary focus of many engineering applications.

The ever-increasing developments in data fusion, sensor technology and high-speed microprocessors made the design in digital domain with high performance. A natural consequence of the progression during the last few decades is the emergence of computationally intelligent system design and development. Artificial neural networks and fuzzy inference systems constitute the core approaches of computational intelligence, whose methods have extensively been used in the applications extending from image pattern recognition to identification and control of nonlinear systems. The reason that lies behind the wideness of this application spectrum is the fact that the architectures of computational intelligence have the capability of perceiving the operating environment and tolerating the faults mostly stemming from the ambiguities in the model of the problem in hand. The verbal power of artificial learning and the numeric power of high-speed connectionist approaches encourage the use of neuro-fuzzy techniques in solving computationally complex problems. A structure with an appropriate learning strategy can teach any mapping to such systems with a predefined realization error bound. The most substantial problem in the use of computationally intelligent architectures is the safety in training. Safety in this context is intimately related to stability, which is the presence of forces driving the parameter update dynamics such that the parameters can not grow unboundedly.

From this point of view, when the widespread applications of neural networks and fuzzy inference systems are taken into consideration in the domain of robotics, the importance of training safety becomes more comprehensible. This tutorial considers various neuro-fuzzy structures and gradient based training procedures.  Consideration is also given to stabilization of training dynamics by utilizing the variable structure systems methodology.

Oct. 24 (TUE) 9:30-12:00
T6) Nanotechnology II
    Prof. U. Hilleringmann, University of Paderborn
    Prof. W. R. Fahrner, University of Hagen

Integration and operation principles of non-electronic micromechanic devices like microtweezers, beryl masks for future lithography, nanometer scale tips of atomic force microscopes and nanofilters for medical applications will be explained. Examples of critical steps to form these devices will be discussed.

Integrated optics on silicon applies typical MOS processing equipment. Depositions of silicon oxide and nitride result in waveguides for visible light. With this it is possible to integrate on-chip interferometers, grids, spectrometers. Examples and concept studies will be presented.

Nanometer scale transistors afford new or enhanced process steps for device integration. Standard manufacturing steps like ion implantation will cause statistical effects, which will limit the device performance in the near future. The aspects of statistical dopant atoms distribution in the channel region of MOS transistors will be discussed regarding statistical measurements on NMOS transistors with channel lengths of only a few tens of nanometer.

Oct. 25 (WED) 9:30-12:00
T7) The Internet, Electronic Commerce, and Factory Automation
    Prof. Alfred C. Weaver, Director
    Internet Technology Innovation Center, University of Virginia

The Internet represents a paradigm shift for the 21st century. Electronic commerce, broadly defined, is the provision of goods and services and information via the Internet. We discuss how this commercial infrastructure can be extended to support real-time factory monitoring and control, which shares the same problems as electronic commerce with regard to real-time data delivery, privacy and security, authentication and authorization, and reliable operation.

This tutorial provides an understanding of the power of the Internet and its potential for worldwide automation. Topics to be discussed include:
*why the Internet is a paradigm shift for its users
*privacy and security via data encryption
*providing real-time access to proprietary data for purposes of monitoring and control, which includes audio, video, sensor data, and control information.
*how the Internet infrastructure developed to support electronic commerce can now be used to support factory automation
*the Java programming language and the Jini networked environment
*predictions of the future.

This tutorial is suitable for engineers, scientists, and managers who are interested in knowing how the Internet can be utilized to provide information reliably and securely and how this existing commercial infrastructure can be utilized for real-time monitoring and control.

Oct. 26 (THU) 9:30-12:00
T8) Soft Computing and its Application
Prof. B. M, 'Dan' Wilamowski, University of Wyoming

Soft computing, known also as computational intelligence, combines neural networks, fuzzy systems, and evolutional computing. All these approaches have already proved their usefulness and they have found many practical applications.

The presentation will start with several examples demonstrating the usefulness of neuro-fuzzy systems, leading to the conclusion that we are at the beginning of the third technological revolution. During the first revolution human labor was replaced by the steam engine.  In the recent second revolution most of clerical and secretary job were replaced by personal computers. Now systems of computational intelligence are capable of replacing highly skilled people with all their experience.

Following this introduction, the concept of the neural networks will be presented, underlining their unique features and limitations. A review and comparison of various supervised and unsupervised learning algorithms will follow. Several special, easy to train, architectures will be shown. The neural network presentation will be illustrated with many practical applications such as speaker identification, sound recognition of various equipment as a diagnosis tool,  written character recognition, data compression using pulse coupled neural networks, time series prediction, etc.

In the next part of the presentation the concept of fuzzy systems, including the conventional Zadeh approach and Takagi-Sugano architecture, will be presented. The basic building blocks of fuzzy systems will be discussed. Comparisons of fuzzy and neural systems will be given and illustrated with several applications. Several new trends in fuzzy systems will be presented. The fuzzy system presentation will be concluded with a demonstration of the fabricated VLSI fuzzy chip, designed by the presenter.

Evolutionally computations, known also as genetic algorithms implements simple methods of search for close to optimal solutions in systems which are often too complicated to be solved by traditional means.

Several hardware implementations of neuro-fuzzy system will be presented, including microprocessors and VLSI. Comparison of accuracy and effectiveness of neural networks and fuzzy systems will be given.

Oct. 27 (FRI) 9:30-12:00
T9) Bond Graph and Related Topics
    Prof. Jean Thoma, University of Waterloo
    Prof. Kazuhiro Tanaka, Kyushu Institute of Technology
    Dr. Katsuya Suzuki, Aichi Science and Technology Foundation

The systems in thermal and chemical engineering consist of various technological components and are governed by many effects of different engineering disciplines. Therefore, for modeling and simulation of such systems we need an unified approach, using the same signs or language for the entire system. The bond graph method with its interdisciplinary outlook seems best suited to the systems. The bond graph approach provides a rigorous and systematic methodology for modeling the multi-energy domain embracing in industrial process engineering the mechanical, electrical, thermal and chemical energies. As opposed to classical modeling methods, bond graph methodology helps to understand the basic physical principles and represent them in a graphical way. Furthermore, the bond graph shows up clearly energy storage and dissipation effects and has the advantage of an efficient transition to the computer by causality assignment and power direction.