Applied Research in Fuzzy Technology INTERNATIONAL SERIES IN INTELLIGENT TECHNOLOGIES Prof. Dr. Dr. h.c. Hans-Jiirgen Zimmermann, Editor Rheinisch-Westfalische Technische Hochschule, Aachen Germany APPLIED RESEARCH IN FUZZY TECHNOLOGY Three years of research at the Laboratory for International Fuzzy Engineering (LIFE), Yokohama, Japan Edited by ANCA L. RALESCU Laboratory for International Fuzzy Engineering Yokohama, Japan and University of Cincinnati, Cincinnati, Ohio, USA .... " Springer Science+Business Media, LLC Library of Congress Cataloging-in-Publication Data Applied research in fuzzy technology : three years of research at the Laboratory for International Fuzzy Engineering (LIFE), Yokohama, Japan / edited by Anca L. Ralescu. p. cm. -- (International series in intelligent technologies) Includes bibliographical references and index. ISBN 978-1-4613-6196-1 ISBN 978-1-4615-2770-1 (eBook) DOI 10.1007/978-1-4615-2770-1 1. Automatic control. 2. Fuzzy logic. 3. Fuzzy systems. 1. Ralescu, Anca L., 1949- II. Laboratory for International Fuzzy Engineering (Yokohama-shi, Japan) III. Series. TJ213.A615 1994 629.8--dc20 94-34459 CIP Copyright © 1994 by Springer Science+Business Media New York Originally published by Kluwer Academic Publishers in 1994 Softcover reprint of the hardcover 1s t edition 1994 AII rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, mechanical, photo-copying, recording, or otherwise, without the prior written permission of the publisher, Springer Science+Business Media, LLC. Printed an acid-free pap er. Contents List of Contributors xi Preface xiii Acknowledgments xvii 1. Future Vision of Fuzzy Engineering Toshiro Terano 2. Decision Support System Satoru Fukami and Minoru Yoneda 17 2. I What is a decision support system ? 17 -A case study of a foreign exchange dealing support expert system - 2. I. 1 Foreign exchange and ambiguity 18 2. I. 2 Basic Concept of FOREX 21 2. 2 Condition recognition in the decision support system 21 2. 2. I Models of the condition recognition mechanism 21 2. 2. 2 Expressing the condition 23 2. 2. 3 Relationship between the condition items 26 2. 2. 4 Updating the condition value 30 2. 3 Fuzzy evaluation and decision making in the decision support system 32 2. 3. I Fuzzy integral 34 2. 3. 2 Fuzzy integral as a multi-attribute utility function 34 2. 3. 3 Formulation of the scenario evaluation 37 2. 3. 4 Interactive method to determine a fuzzy measure 44 2. 3. 5 An example of evaluating a scenario 52 2. 4 Foreign exchange dealing supporting expert system -FOREX- 58 2. 4. I Implementation 59 2. 4. 2 Evaluation trough simulation 61 2. 5 Prospects for the future 63 References 64 vi 3. Intelligent Plant Operation Support Minoru Yoneda and Hiroshi Tsunekawa 67 3. I Functions desired for plant operation support 67 3. 1. I Requirements derived from examples 67 3. 1.2 Functions desired for the operation support system 70 3. 2 Operation support system and background technologies 72 3. 2. I System configuration 72 3. 2. 2 Qualitative process theory 73 3. 2. 3 Problems and solutions in the qualitative process theory 82 3. 3 Simulation by the simplified model 89 3. 3. I Object of simulation 89 3. 3. 2 Reasoning in the qualitative process theory 90 3. 3. 3 Improving the efficiency of reasoning 93 3. 3. 4 State identification 94 3. 3. 5 Examples of the reasoning using the fuzzy theory 95 3. 4 Summary and prospects 100 References 100 4. Fuzzy Modeling and Process Control System Design Kazuyuki Suzuki 103 4. I Subject for process control 103 4. 2 Fuzzy modeling of processes 105 4. 2. I Fuzzy ARX models 106 4. 2. 2 Fuzzy multimodel 112 4.3 Control system design using a fuzzy model 113 4. 3. I Design of a model prediction control system by the fuzzy ARX model 113 4. 3. 2 Design of a multimodel control system using the fuzzy response model 117 4. 4 Application examples 120 4. 4. I Fuzzy model prediction control of a rotary drying incineration furnace for sewage sludge 120 4. 4. 2 Fuzzy multimodel control of a distillation process 129 4.5 Epilogue 135 References 136 vii 5. Inference Function for Understanding Linguistic Instructions Toshihiko Yokogawa 139 5. 1 Linguistic fuzziness and inference 139 5. 1. I Basic sentences for communication 140 5. 1. 2 Analysis offuzziness in natural language 141 5.2 Understanding linguistic instructions using rule based inference 153 5. 2. 1 Internal representation format 154 5. 2. 2 Communication language processing unit 166 5. 2. 3 Inference unit 168 5. 2. 4 Control command generation section 179 5.2.5 Interrupt processing section 180 5. 2. 6 Example of processing 180 5. 2. 7 Summary 186 5. 3 Understanding of linguistic instruction through case-based reasoning 186 5. 3. 1 Introduction of abstraction layers 188 5. 3. 2 Knowledge representation in abstraction layers 189 5. 3. 3 Case-based reasoning method using abstraction layers 193 5. 3. 4 Application to understanding of linguistic instructions 199 5. 3. 5 Conclusion 207 5. 4 Cooperation of case-based reasoning and rule-based reasoning 208 References 213 6. Fuzzy Theory in an Image Understanding Retrieval System Toshio Norita 215 6. 1 Approach to image understanding in LIFE 216 6. 1. 1 Application to facial image understanding and retrieval 218 6. 1. 2 Basis of an image retrieval system 221 6. 1. 3 Model for recognition of facial features 222 6. 2 Recognition of face features based on a recognition model 224 6. 2. 1 Questionnaire 224 6. 2. 2 Detection of physical feature quantities 227 6. 2. 3 Determination of the local degree of certainty 233 6. 2. 4 Determination of the global certainty factor 236 6. 2. 5 Calculation of total certainty factor 237 6. 2. 6 Creation of the data base 238 6. 3 Facial image retrieval system 239 viii 6. 3. I Retrieval by impression words 240 6. 3. 2 Retrieval by sketch 246 6. 3. 3 System configuration 247 6. 4 Conclusion 247 References 250 7. Research into Intelligent Behavior Decision Making of Robots Yoichiro Maeda 253 7. I Ambiguities in Intelligent Robots 254 7.2 An Autonomous Mobile Robot System with Effective Decision Processes for Ambiguous States 254 7.3 Macro Planning Section -Intelligent Path Planning Methods Based on Ambiguous Information - 256 7.3. I Handling of Ambiguities in Path Planning 257 7.3.2 Path Planning Algorithms 259 7. 3. 3 Simulations 263 7. 4 Macro Sensing Part -Hierarchical Sensor Fusion System for Living Object Recognition - 265 7. 4. I Hierarchical Sensor Fusion System 265 7. 4. 2 Recognition Algorithms for the Degree of Certainty of Living Objects 267 7. 4. 3 Actual Experiments 268 7.5 Macro Behavior-Decision Part -Behavior and Decisions Fuzzy Algorithm Tuned according to the Control Purpose - 270 7.5. I Modified Fuzzy Algorithm 272 7.5.2 Ambiguous Concepts and Ambiguous States 274 7.5.3 The Controlled Goal Autonomous Judgment Function 275 7. 5. 4 Behavior-Decision Fuzzy Algorithm 277 7.5.5 Simulation 277 7.6 Robot System for Actual Evaluation 286 7.6. I System Configuration 286 7.6.2 Fuzzy Shell for Intelligent Control 287 7. 6. 3 Autonomous Locomotive Experiments 290 7. 7 Conclusion 292 References 292 ix 8. Fuzzy Neural Net System Toru Yamaguchi, Kenji Goto and Tomohiro Takagi 295 8. 1 What is Fuzzy Neural Net? 296 8. 1. 1 The Potential of Fuzzy neural nets 296 8.1.2 Examples and Classification of Fuzzy Neural net Configurations 301 8. 2 Fuzzy Associative Inference and Fuzzy Knowledge Recursive Learning 307 8. 2. 1 Fuzzy Knowledge Reprsentation Using Associative Memory 307 8.2.2 Fuzzy Knowledge Processing Using Associative Memory 314 8.2.3 Conceptual Fuzzy Sets 318 8.3 Intellectual Interface through Fuzzy Neural net 321 8.3. 1 Intellectual Interface Construction Method Based on Fuzzy Knowledge 321 8. 3. 2 Adaptive Control with Situation Assessment Interface for the Controlled Object 326 8. 3. 3 Human Interface with Beckoning Action 330 8.4 On-Line Learning using Fuzzy Neural Networks 332 8. 4. 1 Learning of Helicopter Flight Operation Knowledge 332 8. 4. 2 Learning a Water Level Prediction Model at a Sewage Treatment Plant 356 8. 5 Future Development of Fuzzy Neural Networks 365 References 366 9. Fuzzy Expert System Shell -LIFE FEShell - Shun'ichi Tano 371 9. 1 Introduction 371 9. 2 System Configuration 372 9. 3 Fuzzy Production System: FPS 373 9.3.1 Outline ofFPS 373 9. 3. 2 Classification of Fuzziness 373 9. 3. 3 Representation of Fuzzy Data and Rules in FPS 375 9. 3. 4 Knowledge Representation 378 9. 3. 5 Method of Inference: Pattern Matching Algorithm 381 9. 4 Fuzzy Frame System: FFS 387 9. 4. 1 Outline of FFS 387 9. 4. 2 Classification of Fuzziness 387 9. 4. 3 Representation Method and Definition Example of a Frame 389 x 9. 4. 4 Inference Method and Example of Frame Processing 391 9. 5 Object Editor. OE 396 9.5. I Outline of OE 396 9.5. 2 Design Policy 396 9. 5. 3.System Configuration and an Example Screen 397 9. 6 Summary and Problems to be Solved in the Future 398 References 399 10. The Fuzzy Computer Hidekazu Tokunaga and Seiji Yasunobu 401 10. I What is a fuzzy computer? 401 10. 2 The architecture of a fuzzy computer prototype system 404 10. 2. I Uncertainty handled by human beings 404 10. 2. 2 Basic processing of fuzziness in fuzzy information processing 405 10. 2. 3 Fuzzy information processing software 408 10. 2. 4 Fuzzy computer prototype hardware 411 10. 3 The fuzzy object-oriented language 416 10. 3. I Effectiveness of object-oriented fuzzy set processing 416 10. 3. 2 Outline of a fuzzy set processing system (FOPS) 418 10. 3. 3 Expression of the fuzzy set 418 10.3.4 Operation offuzzy set using FOPS 422 10.3.5 Performance of FOPS 425 10. 4 The fuzzy computer prototype system 427 10. 4. I Basic operation of fuzzy information processing 428 10. 4. 2 High-speed processing of fuzzy set operations 430 10. 4. 3 Fuzzy Set Processor (FSP) 431 10.4.4 The FUTURE BOARD 437 10. 4. 5 FUTURE BOARD SYSTEM 443 10. 5 Conclusions 449 References 450 Subject Index 451