Rodd / Verbruggen Artificial Intelligence in Real-Time Control 1992

Selected Papers from theIFAC/IFIP/IMACS Symposium, Delft, Netherlands, 16-18 June 1992
1. Auflage 2014
ISBN: 978-1-4832-9902-0
Verlag: Elsevier Science & Techn.
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, 542 Seiten, Web PDF

Reihe: IFAC Symposia Series

The symposium had two main aims, to investigate the state-of-the-art in the application of artificial intelligence techniques in real-time control, and to bring together control system specialists, artificial intelligence specialists and end-users. Many professional engineers working in industry feel that the gap between theory and practice in applying control and systems theory is widening, despite efforts to develop control algorithms. Papers presented at the meeting ranged from the theoretical aspects to the practical applications of artificial intelligence in real-time control. Themes were: the methodology of artificial intelligence techniques in control engineering; the application of artificial intelligence techniques in different areas of control; and hardware and software requirements. This symposium showed that there exist alternative possibilities for control based on artificial intelligence techniques.

Autoren/Hrsg.

Rodd, M. G.

Verbruggen, H. B.

Inhaltsverzeichnis

1;Front Cover;1
2;Artificial Intelligence in Real-Time Control 1992
;4
3;Copyright Page;5
4;Table of Contents;10
5;IFAC SYMPOSIUM ON ARTIFICIAL INTELLIGENCE IN REAL-TIME CONTROL 1992;6
6;PREFACE;8
7;Section I: PLENARY PAPERS;16
7.1;CHAPTER 1. KNOWLEDGE BASED CONTROL: SELECTING THE RIGHT TOOL FOR THE JOB;16
7.1.1;INTRODUCTION;16
7.1.2;APPROPRIATE MODELLING;16
7.1.3;APPROACHES TO MODELLING;17
7.1.4;EXTENDING THE SCOPE OF CONTROL ENGINEERING;20
7.1.5;CONCLUSION;23
7.1.6;REFERENCES;24
7.2;CHAPTER 2. THE FUNCTIONAL LINK NET APPROACH TO THE LEARNING OF REAL-TIME OPTIMAL CONTROL;26
7.2.1;INTRODUCTION;26
7.2.2;IMPLEMENTATION OF ... STEP AHEAD' NEURAL-NET CONTROL;27
7.2.3;LEARNING OPTIMAL OR NEAROPTIMAL CONTROL PATHS;28
7.2.4;REFERENCES;30
8;PART I: THE METHODOLOGY OF ARTIFICIAL INTELLIGENCE TECHNIQUES IN CONTROL SYSTEMS;34
8.1;Section II: Neural Net Control;34
8.1.1;CHAPTER 3. NEURAL NETWORKS APPLIED TO OPTIMAL FLIGHT CONTROL;34
8.1.1.1;1 Introduction;34
8.1.1.2;2 Neural Network as a Multipurpose Nonlinear Function;34
8.1.1.3;3 Aircraft Model;35
8.1.1.4;4 Optimal Control Problems;36
8.1.1.5;5 Conclusions;38
8.1.1.6;6 Acknowledgements;38
8.1.1.7;References;38
8.1.2;CHAPTER 4. ADAPTIVE NEURAL NETWORK CONTROL OF FES-INDUCED CYCLICAL LOWER LEG MOVEMENTS;40
8.1.2.1;INTRODUCTION;40
8.1.2.2;METHODS;41
8.1.2.3;CONCLUSION;44
8.1.2.4;REFERENCES;45
8.1.3;CHAPTER 5. REGULARIZATION AS A SUBSTITUTE FOR PRE-PROCESSING OF DATA IN NEURAL NETWORK TRAINING;46
8.1.3.1;1 Introduction;46
8.1.3.2;2 Neural Networks as adaptive models;47
8.1.3.3;3 Regularization instead of pre-processing;48
8.1.3.4;4 Example;49
8.1.3.5;5 Conclusions;50
8.1.3.6;References;50
8.1.4;CHAPTER 6. NEURAL NETWORK MODELLING AND CONTROL OF A PLANT EXHIBITING THE JUMP PHENOMENA;52
8.1.4.1;INTRODUCTION;52
8.1.4.2;THE JUMP PHENOMENA PLANT;53
8.1.4.3;NEURAL NONLINEAR MODELLING;53
8.1.4.4;TRAINING ALGORITHMS;54
8.1.4.5;NEURAL MODELLING;54
8.1.4.6;NEURAL PREDICTION MODELLING;55
8.1.4.7;NEURAL INTERNAL MODEL CONTROL;55
8.1.4.8;FUTURE WORK;56
8.1.4.9;CONCLUSIONS;56
8.1.4.10;ACKNOWLEDGEMENT;56
8.1.4.11;REFERENCES;56
8.1.5;CHAPTER 7. NEURAL NETWORKS (METHODOLOGIES FOR PROCESS MODELLING AND CONTROL);58
8.1.5.1;ABSTRACT;58
8.1.5.2;INTRODUCTION;58
8.1.5.3;NEURAL NETWORK MODELLING;58
8.1.5.4;SIGMOIDAL FUNCTION NETWORKS;58
8.1.5.5;RADIAL BASIS FUNCTION NETWORKS;59
8.1.5.6;NEURAL NETWORK SOFTWARE SENSORS;60
8.1.5.7;DYNAMIC NEURAL NETWORKS;61
8.1.5.8;NEURAL NETWORK BASED CONTROL;61
8.1.5.9;CONCLUDING REMARKS;62
8.1.5.10;ACKNOWLEDGEMENTS;62
8.1.5.11;REFERENCES;62
8.1.6;CHAPTER 8. PARALLEL NONLINEAR DECOUPLING FOR PROCESS CONTROL - A NARMAX APPROACH;64
8.1.6.1;INTRODUCTION;64
8.1.6.2;NON-LINEAR MODEL REPRESENTATION;64
8.1.6.3;PARALLEL DECOUPLING;65
8.1.6.4;INVERSE-BASED NONLINEAR PROCESS CONTROL;67
8.1.6.5;SIMULATION EXAMPLE - A CHEMICAL REACTOR PROBLEM;67
8.1.6.6;CONCLUSIONS;68
8.1.6.7;ACKNOWLEDGEMENTS;68
8.1.6.8;REFERENCES;68
8.1.7;CHAPTER 9. THE INFLUENCE OF TRAINING DATA SELECTION ON PERFORMANCE OF NEURAL NETWORKS FOR CONTROL OF NON-LINEAR SYSTEMS;70
8.1.7.1;Abstract;70
8.1.7.2;1 Introduction;70
8.1.7.3;2 NNC in Control of Nonlinear Systems;70
8.1.7.4;3 NNC in Control of Systems with Time Delays;73
8.1.7.5;4 Compartive Performance of IMC-PID Controller;74
8.1.7.6;5 Conclusions;75
8.1.7.7;References;75
8.1.8;CHAPTER 10. PROPERTIES OF THE NEURAL NETWORK INTERNAL MODEL CONTROLLER;76
8.1.8.1;INTRODUCTON;76
8.1.8.2;NEURAL NETWORK AS A PREDICTOR;76
8.1.8.3;MODEL .DENTIFICATION;77
8.1.8.4;INTERNAL MODEL CONTROL;78
8.1.8.5;OPTIMAL CONTROLLER DESIGN;78
8.1.8.6;SIMULATON EXAMPLE;79
8.1.8.7;REAL-TIME EXPERIMENTS;80
8.1.8.8;CONCLUSIONS;81
8.1.8.9;REFERENCES;81
8.1.9;CHAPTER 11. A TARGET-DIRECTED NEURALLY CONTROLLED VEHICLE;82
8.1.9.1;INTRODUCTION AND RELATED WORK;82
8.1.9.2;NEURAL TASKS;84
8.1.9.3;CONVENTIONAL ENHANCEMENTS;85
8.1.9.4;EXPERIENCES;85
8.1.9.5;CONCLUSIONS;86
8.1.9.6;REFERENCES;86
8.1.10;CHAPTER 12. EMG PATTERN RECOGNITION BY NEURAL NETWORKS FOR PROSTHETIC FINGERS CONTROL;88
8.1.10.1;INTRODUCTION;88
8.1.10.2;EMG PATTERN RECOGNITION;89
8.1.10.3;CONCLUSION;94
8.1.10.4;ACKNOWLEGEMENTS;94
8.1.10.5;REFERENCES;94
8.1.11;CHAPTER 13. MONITORING AND CONTROL OF POWER SYSTEM VOLTAGE STABILITY USING AN ARTIFICIAL NEURAL NETWORK;96
8.1.11.1;INTRODUCTION;96
8.1.11.2;REVIEW OF MULTIPLE LOAD FLOW SOLUTIONS;97
8.1.11.3;EXAMPLE;100
8.1.11.4;CONCLUSIONS;100
8.1.11.5;REFERENCES;101
8.1.12;CHAPTER 14. MULTI-DIMENSIONAL LOCALLY GENERALIZING NEURAL NETWORKS FOR REAL TIME CONTROL;102
8.1.12.1;INTRODUCTION;102
8.1.12.2;1 Network Review;103
8.1.12.3;2 Network Evaluation On Time Series Prediction;105
8.1.12.4;3 Derivative Estimation;105
8.1.12.5;4 Summary;106
8.1.12.6;References;106
8.2;Section III: Knowledge-based Control;112
8.2.1;CHAPTER 15. INDUCTION OF CONTROL RULES FROM HUMAN SKILL;112
8.2.1.1;Abstract;112
8.2.1.2;1 Introduction;112
8.2.1.3;2 Classification;113
8.2.1.4;3 Induction of Production Rules;114
8.2.1.5;4 Modelling and Control as Classification Problems;115
8.2.1.6;5 Experimental Results - Ball & Beam Control;115
8.2.1.7;References;116
8.2.2;CHAPTER 16. A KNOWLEDGE ACQUISITION AND PROCESSING STRATEGY BASED ON FORMAL CONCEPT ANALYSIS;118
8.2.2.1;INTRODUCTION;118
8.2.2.2;KNOWLEDGE ACQUISITION AND REPRESENTATION;118
8.2.2.3;CONCEPT ANALYSIS;120
8.2.2.4;KNOWLEDGE PROCESSING IN EXPERT SYSTEMS;121
8.2.2.5;DIRECT KNOWLEDGE PROCESSING;122
8.2.2.6;REAL-TIME PROCESS CONTROL;123
8.2.2.7;CONCLUSIONS;123
8.2.2.8;REFERENCES;123
8.2.3;CHAPTER 17. IMPLEMENTATION OF A HUMAN-FRIENDLY SYSTEMS METHODOLOGY FOR INTELLIGENT CONTROL SYSTEM MODELLING AND SIMULATION;124
8.2.3.1;INTRODUCTION;124
8.2.3.2;HUMAN-FRIENDLY SYSTEMS METHODOLOGY;125
8.2.3.3;MODELING AND SIMULATION SUPPORT;126
8.2.3.4;INTELLIGENT PROCESS CONTROL;126
8.2.3.5;CONCLUSION;128
8.2.3.6;REFERENCES;129
8.2.4;CHAPTER 18. MACHINE LEARNING USING VERSION SPACES FOR A POWER DISTRIBUTION NETWORK FAULT DIAGNOSTICIAN;130
8.2.4.1;INTRODUCTION;130
8.2.4.2;VERSION SPACES;131
8.2.4.3;TESTING PROCEDURE;132
8.2.4.4;RESULTS;133
8.2.4.5;DISCUSSION;134
8.2.4.6;CONCLUSION;135
8.2.4.7;ACKNOWLEDGEMENTS;135
8.2.4.8;REFERENCES;135
8.2.5;CHAPTER 19. STABILITY OF FUZZY CONTROL SYSTEMS BY USING NONLINEAR SYSTEM THEORY;136
8.2.5.1;INTRODUCTION;136
8.2.5.2;FUZZY CONTROL AS A NONLINEAR CONTROL SYSTEM;137
8.2.5.3;PRELIMINARY ANALYSIS OF A FUZZY CONTROL SYSTEM;137
8.2.5.4;STABILITY AND ROBUSTNESS INDICES;138
8.2.5.5;INPUT-OUTPUT STABILITY;139
8.2.5.6;CONCLUSIONS;141
8.2.5.7;REFERENCES;141
8.2.6;CHAPTER 20. STABILITY OF FEEDBACK SYSTEMS WITH UNCERTAIN DYNAMICS;142
8.2.6.1;1. INTRODUCNON;142
8.2.6.2;2. INPUT-OUTPUT STABILITY;142
8.2.6.3;3. APPLICATON OF THE CONICITY CRITERION;144
8.2.6.4;4. UNCERTAIN DYNAMICS;147
8.2.6.5;5. CONCLUSIONS;147
8.2.6.6;ACKNOWLEDGEMENT;147
8.2.6.7;REFERENCES;147
8.2.7;CHAPTER 21. A COMPUTATIONAL CAUSAL MODEL FOR PROCESS SUPERVISION;148
8.2.7.1;1 Introduction;148
8.2.7.2;2 Representing process knowledge;149
8.2.7.3;3 Modeling with q-automata;151
8.2.7.4;4 Qualitative automata dynamics;153
8.2.7.5;5 Prediction algorithm and sprinkler example;153
8.2.7.6;6 Conclusion and perspectives;153
8.2.8;CHAPTER 22. DESIGN OF AN INTELLIGENT SUPERVISOR OF A SHIP ENGINE ROOM;156
8.2.8.1;INTRODUCTION;156
8.2.8.2;MODEL DESCRIPTION;157
8.2.8.3;SUPERVISOR STRUCTURE;158
8.2.8.4;APPLICATION EXAMPLE;159
8.2.8.5;CONCLUSIONS;161
8.2.8.6;ACKNOWLEDGEMENT;161
8.2.8.7;REFERENCES;161
8.2.9;CHAPTER 23. QUALITATIVE MODELLING AND SIMULATION BY PIECEWISE LINEAR ANALYSIS;162
8.2.9.1;1 Introduction;162
8.2.9.2;2 Qualitative Model and Analysis;162
8.2.9.3;3 Piecewise Linear Functions;163
8.2.9.4;4 The construction of the family of piecewise linear dynamical systems;164
8.2.9.5;5 Qualitative Analysis of Dynamical Systems;165
8.2.9.6;6 Applications to the example;165
8.2.9.7;7 Conclusions;166
8.2.9.8;8 References;166
8.2.9.9;9 Appendix;166
8.2.10;CHAPTER 24. IMPLEMENTATION OF A KNOWLEDGE-BASED PID AUTO-TUNER;168
8.2.10.1;Abstract;168
8.2.10.2;1 Introduction;168
8.2.10.3;2 Hardware;168
8.2.10.4;3 Software;168
8.2.10.5;4 Knowledge Based Design;169
8.2.10.6;5 Knowledge Sources;169
8.2.10.7;6 A Sample Session;170
8.2.10.8;7 Conclusions;171
8.2.10.9;8 References;171
8.2.11;CHAPTER 25. DIMENSIONS OF LEARNING IN A REAL-TIME KNOWLEDGE-BASED CONTROL SYSTEM;174
8.2.11.1;INTRODUCTION;174
8.2.11.2;THE APPROACH;174
8.2.11.3;ON THE CONTROL STRATEGIES;175
8.2.11.4;THE INFORMATION COMMUNICATED BETWEEN CONTROLLERS;175
8.2.11.5;THE RULES AND META-RULES TO CONTROL THE TRAFFIC LIGHTS;175
8.2.11.6;ON SCENARIO GENERATION;176
8.2.11.7;THE OPTIMIZATION OF THE RULE BASE;176
8.2.11.8;THE IMPLEMENTATION OF A PROTOTYPE SYSTEM;177
8.2.11.9;SUMMARY;178
8.2.11.10;REFERENCES;178
8.2.12;CHAPTER 26. EDUCATIONAL ASPECT OF EXPERT CONTROL OF TECHNOLOGICAL PROCESSES;180
8.2.12.1;INTRODUCTION;180
8.2.12.2;EXPERT CONTROL;180
8.2.12.3;TUTORIAL ENVIRONMENT FOR EXPERT CONTROL;181
8.2.12.4;EXAMPLE;183
8.2.12.5;CONCLUSION;183
8.2.12.6;REFERENCES;184
8.2.13;CHAPTER 27. A PREDICTABLE REAL-TIME EXPERT SYSTEM FOR MULTI-SENSOR FUSION;186
8.2.13.1;INTRODUCTION;186
8.2.13.2;TERMINOLOGY;186
8.2.13.3;DEFINITION;187
8.2.13.4;A BASIC THEOREM;187
8.2.13.5;PREDICTABILITY OF MAXIMUM RESPONSE TIMES;188
8.2.13.6;CONCLUSION;190
8.2.13.7;REFERENCES;190
8.3;Section IV: Fuzzy Control;192
8.3.1;CHAPTER 28. ADAPTIVE AND SUPPLEMENTARY INTELLIGENT CONTROL OF POWER SYSTEM STABILIZERS;192
8.3.1.1;1. INTRODUCTION;192
8.3.1.2;2. THE POWER SYSTEM STABILIZER;193
8.3.1.3;3. THE EXPERT SYSTEM;194
8.3.1.4;4. APPLICATION OF CASE STUDY;196
8.3.1.5;5. CONCLUSION;196
8.3.2;CHAPTER 29. FUZZY INFERENCE IN RULE-BASED REAL-TIME CONTROL;198
8.3.2.1;1 Introduction;198
8.3.2.2;2 Fuzzy rules and inference;199
8.3.2.3;3 Fuzzy inference engine;199
8.3.2.4;4 An example;201
8.3.2.5;5 Conclusions;202
8.3.2.6;References;203
8.3.3;CHAPTER 30. LABORATORY EVALUATION OF FUZZY CONTROLLERS;204
8.3.3.1;INTRODUCTION;204
8.3.3.2;FUZZY CONTROL;204
8.3.3.3;PID CONTROLLER;205
8.3.3.4;LQ CONTROLLER;205
8.3.3.5;FUZZY TUNED FID CONTROLLER;206
8.3.3.6;BALL & FLATE LABORATORY MODEL;206
8.3.3.7;SOFTWARE SUPPORT;207
8.3.3.8;EXPERIMENTS;207
8.3.3.9;CONCLUSION;208
8.3.3.10;REFERENCES;209
8.4;Section V: Monitoring and Fault Diagnosis;210
8.4.1;CHAPTER 31. SUPERVISORY CONTROL OF MODE-SWITCH PROCESSES: APPLICATION TO A FLEXIBLE BEAM;210
8.4.1.1;1. INTRODUCTION;210
8.4.1.2;2. FLEXIBLE BEAM;211
8.4.1.3;3. SUPERVISORY CONTROL;211
8.4.1.4;4. MODE DETECTOR;212
8.4.1.5;5. EXPERIMENTS AND RESULTS;213
8.4.1.6;6. CONCLUSIONS;214
8.4.1.7;REFERENCES;215
8.4.2;CHAPTER 32. KNOWLEDGE SPECIFICATION AND REPRESENTATION FOR AN "INTELLIGENT" INTERFACE DEVOTED TO PROCESS MONITORING AND SUPERVISION;216
8.4.2.1;INTRODUCTION;216
8.4.2.2;THE DECISIONAL MODULE OF IMAGERY;216
8.4.2.3;THE OBJECTS MANIPULATED BY THE "INTELLIGENT" INTERFACE;217
8.4.2.4;KNOWLEDGE SPECIFICATION AND REPRESENTATION FOR THE "INTELLIGENT" INTERFACE;218
8.4.2.5;CONCLUSION;220
8.4.2.6;REFERENCES;221
8.4.2.7;APPENDICES;221
8.4.3;CHAPTER 33. INTEGRATION OF CONTROL AND ERROR MANAGEMENT FOR A FLEXIBLE ASSEMBLY CELL, USING A COST FUNCTION;222
8.4.3.1;I Introduction;222
8.4.3.2;II Planning and Control for flexible assembly;222
8.4.3.3;III A reference model for planning and control of flexible assembly systems;225
8.4.3.4;.V Scheduling and a cost function for the selection of strategies for primitive operations;226
8.4.3.5;V Conclusions and future work;227
8.4.3.6;Acknowledgements;227
8.4.3.7;Literature;227
8.4.4;CHAPTER 34. PRINCIPLES OF MODEL-BASED FAULT DETECTION;228
8.4.4.1;Introduction;228
8.4.4.2;General Concept of Analytical Model-based FDI;229
8.4.4.3;Principles of Residual Generation;230
8.4.4.4;Parameter Identification Approach;230
8.4.4.5;Observer-based Residual Generation;230
8.4.4.6;Generation of structured residuals;231
8.4.4.7;Innovation Test;231
8.4.4.8;Fault Detection Filter (FDF);231
8.4.4.9;Parity Space Approach;232
8.4.4.10;Unknown Input Observer Scheme (UIOS);232
8.4.4.11;Multiple Hypotheses Tests;233
8.4.4.12;Combined Hardware-Software Redundancy;233
8.4.4.13;Hierarchical Observer Scheme (HOS);233
8.4.4.14;Decision and Monitoring;234
8.4.4.15;Conclusions;234
8.4.4.16;References;235
8.4.5;CHAPTER 35. ON-LINE RESIDUAL COMPENSATION IN ROBUST FAULT DIAGNOSIS OF DYNAMIC SYSTEMS;236
8.4.5.1;1 Introduction;236
8.4.5.2;2 Basic concepts of residual generation;237
8.4.5.3;3 On-line residual compensation;239
8.4.5.4;4 An example: Robust detection of Incipient faults in jet engine sensors;240
8.4.5.5;5 Conclusion;242
8.4.5.6;6 Acknowledgements;242
8.4.5.7;7 Reference;242
8.4.6;CHAPTER 36. GEOMETRIC TOOLS FOR AN OBSERVER-BASED APPROACH TO RESIDUAL GENERATION;244
8.4.6.1;1 INTRODUCTION;244
8.4.6.2;2 RESIDUAL GENERATION;244
8.4.6.3;3 PHYSICAL APPLICATION;246
8.4.6.4;REFERENCES;248
8.4.7;CHAPTER 37. EXAMPLES FOR FAULT DETECTION IN CLOSED LOOPS;250
8.4.7.1;1. Introduction;250
8.4.7.2;2. Models for the closed loop and its components;250
8.4.7.3;3. Fault models;251
8.4.7.4;4. Fault detection in control loops;251
8.4.7.5;5. Simulation results with state estimation;253
8.4.7.6;6. Conclusions;255
8.4.7.7;7. References;255
8.4.8;CHAPTER 38. A METHOD FOR FAULT DETECTION USING PARAMETER AND STATE ESTIMATION;256
8.4.8.1;1. Introduction;256
8.4.8.2;2. Method;256
8.4.8.3;3. Identification;257
8.4.8.4;4. State Variable Filter;257
8.4.8.5;5. Realization of the SVF by digital computers;258
8.4.8.6;6. Examples;259
8.4.8.7;7. Conclusion;260
8.4.8.8;References;260
8.4.9;CHAPTER 39. SUPERVISION AND CONTROL OF AN EXOTHERMIC BATCH PROCESS;262
8.4.9.1;Introduction;262
8.4.9.2;The Problem;262
8.4.9.3;The New Control Concept;263
8.4.9.4;Early Detection of a Hazardous Secondary Reaction;265
8.4.9.5;Conclusion;266
8.4.9.6;References;266
8.4.10;CHAPTER 40. MARKOVIAN RELIABILITY ANALYSIS OF STATE-ESTIMATOR-BASED INSTRUMENT FAULT DETECTION SCHEMES;268
8.4.10.1;INTRODUCTION;268
8.4.10.2;DESCRIPTION OF ESTIMATOR SCHEMES;268
8.4.10.3;BASIC THEORY OF MARKOV PROCESSES;269
8.4.10.4;MODELING OF ESTIMATOR SCHEMES;270
8.4.10.5;COMPARISON OF ESTIMATOR SCHEMES;271
8.4.10.6;CONCLUSIONS;273
8.4.10.7;REFERENCES;273
8.4.11;CHAPTER 41. NEURAL NETWORK MODELS AND STATISTICAL TESTS AS FLEXIBLE BASE FOR INTELLIGENT FAULT DIAGNOSIS;274
8.4.11.1;1. INTRODUCTION;274
8.4.11.2;2. CONVERSION TO FLEXIBLE FAULT DIAGNOSIS TECHNIQUES;274
8.4.11.3;3. QUASI - NEURAL ANALYTICAL KNOWLEDGE POR THE INTELLIGENT FAULT DIAGNOSIS SYSTEMS;276
8.4.11.4;4. A BETTER INSIGHT INTO THE QUASI-NEURAL PAUL DIAGNOSIS "PENT-HOUSE";277
8.4.11.5;5. EXAMPLES;278
8.4.11.6;6. CONCLUSION;278
8.4.12;CHAPTER 42. MULTIVALUED LOGIC VOTING SCHEME FOR RESIDUAL EVALUATION IN FAILURE DETECTION AND ISOLATION SYSTEMS;282
8.4.12.1;I - INTRODUCTION;282
8.4.12.2;II - SYSTEM DESCRIPTION;283
8.4.12.3;Ill - DETECTION PROCEDURE;283
8.4.12.4;IV - BINARY LOGIC CASE;284
8.4.12.5;V - .UTLTIVALUED LOGIC CASE;285
8.4.12.6;VI. DISCUSSION;286
8.4.12.7;VII. CONCLUSION;287
8.4.12.8;REFERENCES ;287
8.4.13;CHAPTER 43. KNOWLEDGE-BASED DIAGNOSIS IN INFORMATION POOR PLANTS: A MATERIALS ACCOUNTANCY APPLICATION;288
8.4.13.1;INTRODUCTON;288
8.4.13.2;BACKGROUND;288
8.4.13.3;DIAGNOSIS IN NRTMA;290
8.4.13.4;AN EXAMPLE;292
8.4.13.5;CONCLUSIONS;293
8.4.13.6;REFERENCES;293
8.4.14;CHAPTER 44. LOGIC-BASED PROCESS DIAGNOSIS UTILISING THE CAUSAL STRUCTURE OF DYNAMICAL SYSTEMS;294
8.4.14.1;INTRODUCTION;294
8.4.14.2;1. THE DIAGNOSIS PROBLEM;295
8.4.14.3;2. THE ASSERTIONAL-LOGIC DESCRIPTION OF THE PROCESS;295
8.4.14.4;3. STATEMENT OF THE DIAGMOSIS PROBLEM IN ASSERTIONAL LOGIC;296
8.4.14.5;4. DIRECT SOLUTION OF THE DIAGNOSIS PROBLEM BY MEANS OF THE RESOLUTION METHOD;296
8.4.14.6;5. THE CAUSAL STRUCTURE OF DYNAMICAL SYSTEMS;296
8.4.14.7;6. A DECOMPOSITION PRINCIPLE FOR THE DIAGNOSIS PROBLEM;297
8.4.14.8;7 . THE DIAGNOSIS SYSTEM;298
8.4.14.9;8. EXAMPLE;298
8.4.14.10;REFERENCES;299
8.4.15;CHAPTER 45. KNOWLEDGE BASED SENSOR FAULT DETECTION FOR GAS TURBINES UNDER CONSIDERATION OF MODEL BASED METHODS;302
8.4.15.1;INTRODUCTION;302
8.4.15.2;MODELING;302
8.4.15.3;KNOWLEDGE BASED METHODS;303
8.4.15.4;THE EXPERT SYSTEM XSD;306
8.4.15.5;USE OF UNCERTAIN LOGIC;306
8.4.15.6;CONCLUSIONS AND OUTLOOK;307
8.4.15.7;REFERENCES;307
8.4.16;CHAPTER 46. TOWARDS A GENERAL MULTI-MODEL-BASED METHODOLOGY FOR DIAGNOSIS SYSTEMS;308
8.4.16.1;ABSTRACT;308
8.4.16.2;KEYWORDS;308
8.4.16.3;A PETRI NET BASED MODEL FOR FAULT DIAGNOSIS;308
8.4.16.4;A FAULT ISOLATION REASONING;309
8.4.16.5;INTRODUCTION;308
8.4.16.6;TOWARDS A MUm-MODEL-BASED METHODOLOGY;310
8.4.16.7;CONCLUSION;310
8.4.16.8;REFERENCES;311
8.4.17;CHAPTER 47. AN ADAPTIVE DECISION SYSTEM USING PATTERN RECOGNITION;314
8.4.17.1;INTRODUCTION;314
8.4.17.2;DIAGNOSIS AND PATTERN RECOGNITION;314
8.4.17.3;USUAL PROBLEMS IN DIAGNOSIS;315
8.4.17.4;DECISION WITH REJECT;315
8.4.17.5;IMPROVEMENT OF THE DIAGNOSTIC SYSTEM;316
8.4.17.6;CONCLUSION;317
8.4.17.7;REFERENCES;317
8.5;Section VI: Genetic Algorithms and Learning;318
8.5.1;CHAPTER 48. INTELLIGENT REAL-TIME CONTROL OF A MULTIFINGERED ROBOT GRIPPER BY LEARNING INCREMENTAL ACTIONS;318
8.5.1.1;INTRODUCTION;318
8.5.1.2;BASICS;319
8.5.1.3;AGRICOLA;320
8.5.1.4;CONCLUSION;323
8.5.1.5;REFERENCES;323
8.5.2;CHAPTER 49. SYNTHESIS OF OPTIMAL CONTROL USING NEURAL NETWORK WITH MIXED STRUCTURE;326
8.5.2.1;INTRODUCTION;326
8.5.2.2;NEURAL NETWORK WITH MIXED STRUCTURE;326
8.5.2.3;LEARNING ALGORITHM;327
8.5.2.4;SYNTHESIS OF OPTIMAL CONTROL;328
8.5.2.5;ILLUSTRATIVE EXAMPLE;329
8.5.2.6;CONCLUSIONS;329
8.5.2.7;REFERENCES;329
8.5.3;CHAPTER 50. LEARNING TO AVOID COLLISIONS: A REINFORCEMENT LEARNING PARADIGM FOR MOBILE ROBOT NAVIGATION;332
8.5.3.1;INTRODUCTION;332
8.5.3.2;SELF-ORGANISING STATE-SPACE QUANTISATION;333
8.5.3.3;LEARNING THE CORRECT ACTIONS;333
8.5.3.4;EXPERIMENTS AND RESULTS;334
8.5.3.5;CONCLUSIONS;335
8.5.3.6;REFERENCES;336
8.5.4;CHAPTER 51. GENETIC ALGORITHMS FOR PROCESS CONTROL: A SURVEY;338
8.5.4.1;INTRODUCTION;338
8.5.4.2;BRIEF DESGRKTION OF GAs;338
8.5.4.3;OFF-LINE CONTROL;339
8.5.4.4;ON-UNE CONTROL;341
8.5.4.5;DISCUSSION AND CONCLUSION;342
8.5.4.6;REFERENCES;343
8.5.5;CHAPTER 52. AN ADAPTIVE SYSTEM FOR PROCESS CONTROL USING GENETIC ALGORITHMS;344
8.5.5.1;INTRODUCTION;344
8.5.5.2;PROBLEM ENVIRONMENT;345
8.5.5.3;STRUCTURE OP THE ADAPTIVE CONTROLLER;345
8.5.5.4;SUMMARY;349
8.5.5.5;REFERENCES;349
8.5.6;CHAPTER 53. REAL-TIME ACQUISITION OF FUZZY RULES USING GENETIC ALGORITHMS;350
8.5.6.1;1 INTRODUCTION;350
8.5.6.2;2 GENETICAL GORITHMS: A REVIEW;351
8.5.6.3;3 A GENETIC ALGORITHM FOR FUZZY LOGIC CONTROL;351
8.5.6.4;4 THE TASK ENVIRONMENT;352
8.5.6.5;5 PERFORMANCE EVALUATION;352
8.5.6.6;6 ACQUISITION OF RULES;352
8.5.6.7;7 SIMULATION RESULTS;353
8.5.6.8;8 CONCLUSION;353
8.5.6.9;References;353
8.5.7;CHAPTER 54. AUTOMATED SYNTHESIS OF CONTROL FOR NONLINEAR DYNAMIC SYSTEMS;356
8.5.7.1;INTRODUCTION;356
8.5.7.2;CASE STUDY 1;357
8.5.7.3;CASE STUDY 2;359
8.5.7.4;CONCLUSION;361
8.5.7.5;REFERENCES;361
8.6;Section VII: Qualitative Reasoning;362
8.6.1;CHAPTER 55. ON REPRESENTATIONS FOR CONTINUOUS DYNAMIC SYSTEMS;362
8.6.1.1;INTRODUCTION;362
8.6.1.2;THE EXAMPLE SYSTEM;363
8.6.1.3;A RULE BASED APPROACH;363
8.6.1.4;A QUALITATIVE REASONING APPROACH;364
8.6.1.5;A HYBRID APPROACH;365
8.6.1.6;DISCUSSION AND CONCLUSION;366
8.6.1.7;ACKNOWLEDGMENT;367
8.6.1.8;REFERENCES;367
8.6.2;CHAPTER 56. PROCESS KNOWLEDGE ACQUISITION AND CONTROL BY QUANTITATIVE AND QUALITATIVE COMPLEMENTARITY;368
8.6.2.1;1. Introduction;368
8.6.2.2;2. Model-based knowledge acquisition;368
8.6.2.3;3. Inference model building;370
8.6.2.4;4. Preliminary study of qualitative control for the cement rotary kiln process;370
8.6.2.5;5. Conclusion;372
8.6.2.6;APPENDIX;373
8.6.2.7;REFERENCE;373
8.6.3;CHAPTER 57. MODEL-BASED DIAGNOSIS - STATE TRANSITION EVENTS AND CONSTRAINT EQUATIONS;374
8.6.3.1;1· Introduction;374
8.6.3.2;2. DMP;374
8.6.3.3;3. MIDAS;376
8.6.3.4;4· Steritherm diagnosis;378
8.6.3.5;5. Conclusions;379
8.6.3.6;References;379
8.6.4;CHAPTER 58. MULTIPLE MODELS BASED ON FUZZY QUALITATIVE MODELLING;380
8.6.4.1;INTRODUCTION;380
8.6.4.2;AN OVERVIEW OF FUZZY QUALITATIVE SIMULATION;380
8.6.4.3;MULTIPLE MODELS BASED ON MODELLING DIMENSIONS;381
8.6.4.4;IMPLICATIONS OF MULTIPLE MODELS FOR MODEL-BASED REASONING;382
8.6.4.5;EXAMPLE;383
8.6.4.6;CONCLUSION;385
8.6.4.7;REFERENCES;385
9;PART II: THE APPLICATION OF ARTIFICIAL INTELLIGENCE TECHNIQUES IN DIFFERENT AREAS OF CONTROL;386
9.1;Section VIII: Process Control;386
9.1.1;CHAPTER 59. ARCHITECTURES AND TECHNIQUES OF ARTIFICIAL INTELLIGENCE IN PROCESS CONTROL;386
9.1.1.1;INTRODUCTION;386
9.1.1.2;AI ARCHITECTURES IN PROCESS CONTROL;387
9.1.1.3;AI TECHNIQUES IN PROCESS CONTROL;389
9.1.1.4;THE ROLE OF THE OPERATOR;389
9.1.1.5;CONCLUDING REMARKS;390
9.1.1.6;REFERENCES;391
9.1.2;CHAPTER 60. REALTIME SUPERVISORY CONTROL FOR INDUSTRIAL PROCESSES;392
9.1.2.1;INTRODUCTION;392
9.1.2.2;KNOWLEDGE-BASED CONTROL;392
9.1.2.3;AN EXAMPLE: LIQUID LEVEL RIG;394
9.1.2.4;SYSTEM IMPLEMENTATION;394
9.1.2.5;EXPERIMENTAL RESULTS;395
9.1.2.6;CONCLUSIONS;396
9.1.2.7;REFERENCES;396
9.1.3;CHAPTER 61. THE DEVELOPMENT OF AN INTELLIGENT MONITORING AND CONTROL SYSTEM FOR A SOLVENT EXTRACTION PROCESS;398
9.1.3.1;INTRODUCTION;398
9.1.3.2;REAL-TI.. ATTRIBUTES;399
9.1.3.3;PROCESS OVERVIEW;400
9.1.3.4;IMPLEMENTATION;400
9.1.3.5;APPLICATION ARCHITECTURE;401
9.1.3.6;INTELLIGENT CONTROL;401
9.1.3.7;SUMMARY;402
9.1.3.8;REFERENCES;402
9.1.4;CHAPTER 62. KNOWLEDGE-BASED SYSTEMS FOR REAL-TIME PROCESS CONTROL: THE MIP PROJECT;406
9.1.4.1;INTRODUCTION;406
9.1.4.2;STATEMENT OF THE PROBLEM;406
9.1.4.3;HARDWARE ARCHITECTURE;407
9.1.4.4;KNOWLEDGE ARCHITECTURE;407
9.1.4.5;SOFTWARE ARCHITECTURE;408
9.1.4.6;METHODOLOGY ISSUES;409
9.1.4.7;CONCLUSIONS;410
9.1.4.8;ACKNOWLEDGEMENTS;411
9.1.4.9;REFERENCES;411
9.1.5;CHAPTER 63. USING GRATE TO BUILD COOPERATING AGENTS FOR INDUSTRIAL CONTROL;412
9.1.5.1;INTRODUCTION;412
9.1.5.2;GRATE ARCHITECTURE;413
9.1.5.3;BUILDING GRATE APPLICATIONS;414
9.1.5.4;GRATE IN INDUSTRIAL CONTROL;416
9.1.5.5;CONCLUSIONS;416
9.1.5.6;ACKNOWLEDGMENTS;417
9.1.5.7;REFERENCES;417
9.1.6;CHAPTER 64. INTELLIGENT TUNING OF P+I CONTROLLERS FOR BIOPROCESS APPLICATION;418
9.1.6.1;ABSTRACT;418
9.1.6.2;KEYWORDS;418
9.1.6.3;INTRODUOTON;418
9.1.6.4;NEURAL NETWORK BASED CONTROL;419
9.1.6.5;OPTIMAL PIP CONTROL;420
9.1.6.6;BIOPROCESS MODEL;420
9.1.6.7;CONTROL RESULTS;421
9.1.6.8;CONCLUSIONS;421
9.1.6.9;REFERENCES;422
9.1.7;CHAPTER 65. FAULT DETECTION AND EMERGENCY CONTROL IN POWER SYSTEMS;424
9.1.7.1;INTRODUCTION;424
9.1.7.2;RELATED WORK;425
9.1.7.3;KNOWLEDGE BASED SYSTEM DESIGN;425
9.1.7.4;CONCLUSIONS;429
9.1.7.5;REFERENCES;429
9.1.8;CHAPTER 66. MODEL-BASED DIAGNOSIS FOR CONTINUOUS PROCESS SUPERVISION: THE ALEXIP EXPERIENCE;430
9.1.8.1;INTRODUCTION;430
9.1.8.2;AVAILABLE PROCESS INFORMATON;430
9.1.8.3;REPRESENTING INTERACTONS BETWEEN VARLIABLES;431
9.1.8.4;DETERMINING DOMINANT AND MASKED EVENTS;432
9.1.8.5;DEALING WITH TRANSIENT STATES;433
9.1.8.6;DETERMINING THE REFERENCE STATE;434
9.1.8.7;DISCUSSION;435
9.1.8.8;ACKNOWLEDGEMENTS;435
9.1.8.9;REFERENCES;435
9.1.9;CHAPTER 67. REACTIVE PROCESS CONTROL USING A BLACKBOARD ARCHITECTURE;438
9.1.9.1;1. INTRODUCTION;438
9.1.9.2;2. RELATED WORK;439
9.1.9.3;3. REQUIREMENTS FOR INTELLIGENT PROCESS CONTROL;442
9.1.9.4;4. PROCESS CONTROL IN A REAL-TIME BLACKBOARD FRAMEWORK;442
9.1.9.5;5. CONCLUSIONS;444
9.1.9.6;6. FURTHER WORK;444
9.1.9.7;7. REFERENCES;444
9.2;Section IX: Biotechnology;446
9.2.1;CHAPTER 68. PATTERN RECOGNITION FOR BIOPROCESS CONTROL;446
9.2.1.1;PATTERNS;446
9.2.1.2;INTRODUCTON;446
9.2.1.3;EXPERIMENTAL EXPERIENCE;447
9.2.2;CHAPTER 69. ENHANCING FERMENTATION DEVELOPMENT PROCEDURES VIA ARTIFICIAL NEURAL NETWORKS;450
9.2.2.1;ABSTRACT;450
9.2.2.2;KEYWORDS;450
9.2.2.3;INTRODUCTION;450
9.2.2.4;FEEDFORWARD NEURAL NETWORKS ARTIFICIAL;451
9.2.2.5;DYNAMIC NEURAL NETWORKS;452
9.2.2.6;ON-LINE BIOPROCESS VARIABLE ESTIMATION;453
9.2.2.7;EXPERIMENTAL DESIGN;453
9.2.2.8;DISCUSSION;454
9.2.2.9;ACKNOWLEDGEMENTS;454
9.2.2.10;REFERENCES;454
9.2.3;CHAPTER 70. ARTIFICIAL INTELLIGENCE IN THE CONTROL OF A CLASS OF FERMENTATION PROCESSES;456
9.2.3.1;INTRODUCTION;456
9.2.3.2;THE AUTOREGRESSIVE (AR) MODEL FOR THE FERMENTATION PROCESS;456
9.2.3.3;NONLINEAR MODEL OF THE PROCESS;457
9.2.3.4;SELF ORGANIZING FUZZY LOGIC CONTROLLER;457
9.2.3.5;PATTERN RECOGNITION FOR MODELLING AND CONTROL OF FERMENTA.ON PROCESS;458
9.2.3.6;ARTIFICIAL NEURAL NETWORKS;458
9.2.3.7;THE INTELLIGENT SYSTEM APPROACH;459
9.2.3.8;CONCLUSION;459
9.2.3.9;REFERENCES;460
9.2.4;CHAPTER 71. A TASK DECOMPOSITION APPROACH TO USING NEURAL NETWORKS FOR THE INTERPRETATION OF BIOPROCESS DATA;462
9.2.4.1;Introduction;462
9.2.4.2;Materials and Metliods;462
9.2.4.3;Blackbox Approach;463
9.2.4.4;Results;463
9.2.4.5;Task decomposition Approach;463
9.2.4.6;Results;464
9.2.4.7;Pruning;464
9.2.4.8;Conclusions;464
9.2.4.9;Literature;464
10;PART III: HARDWARE AND SOFTWARE REQUIREMENTS;468
10.1;Section X: Temporal Reasoning;468
10.1.1;CHAPTER 72. TOP-DOWN DESIGN OF EMBEDDED REAL-TIME AI SYSTEMS;468
10.1.1.1;INTRODUCTION;468
10.1.1.2;PROGRAMMING LANGUAGE;468
10.1.1.3;SPECIFICATIONS;469
10.1.1.4;EXAMPLE RAILWAY CROSSING;470
10.1.1.5;DECOMPOSITION OF CONTROL;471
10.1.1.6;VERIFICATION;472
10.1.1.7;IMPLEMENTATION OF WAY;472
10.1.1.8;REFERENCES;473
10.1.2;CHAPTER 73. A TEMPORAL BLACKBOARD STRUCTURE FOR PROCESS CONTROL;474
10.1.2.1;INTRODUCTION;474
10.1.2.2;TEMPORAL INFORMATION;475
10.1.2.3;TEMPORAL MODEL;475
10.1.2.4;TEMPORAL MANAGEMENT;476
10.1.2.5;BLACKBOARD STRUCTURE;477
10.1.2.6;EXAMPLE;479
10.1.2.7;CONCLUSIONS;480
10.1.2.8;REFERENCES;480
10.2;Section XI: New Paradigms for Real-time Control;482
10.2.1;CHAPTER 74. REINFORCEMENT LEARNING AND RECRUITMENT MECHANISM FOR ADAPTIVE DISTRIBUTED CONTROL;482
10.2.1.1;INTRODUCnON;482
10.2.1.2;THE GENERAL APPROACH;483
10.2.1.3;Q_LEARNING;484
10.2.1.4;THE RECRUITMENT MECHANISM;486
10.2.1.5;EXPERIMENTAL RESULTS;487
10.2.1.6;CONCLUSIONS;487
10.2.1.7;REFERENCES;488
10.2.2;CHAPTER 75. USING NEURAL-NET COMPUTING TO FORMULATE REAL-TIME CONTROL STRATEGIES;490
10.2.2.1;INTRODUCTION;490
10.2.2.2;THE LEARNING OF CONTROL ACTIONS: CONCEPT IDENTIHC ATION;491
10.2.2.3;DISCUSSION OF THE METHOD;491
10.2.2.4;REFERENCES;492
10.3;Section XII: Real-time Environments for Intelligent Control;498
10.3.1;CHAPTER 76. A SURVEY OF COMMERCIAL REAL-TIME EXPERT SYSTEM ENVIRONMENTS;498
10.3.1.1;1. Introduction;498
10.3.1.2;2. Real-time aspects;499
10.3.1.3;3. Functionality;499
10.3.1.4;4. Tool survey;501
10.3.1.5;5. Summary;505
10.3.1.6;References;505
10.3.2;CHAPTER 77. DESIGNING REAL-TIME KNOWLEDGE BASED SYSTEMS WITH PERFECT;506
10.3.2.1;INTRODUCTION;506
10.3.2.2;DESIGNING THE COOPERATION OF THE RTKBS WITH EXISTING SOFTWARE SYSTEMS;507
10.3.2.3;THE MODELING TECHNIQUE OF PERFECT;507
10.3.2.4;THE MONITOR AND DIAGNOSE FUNCTION OF THE RTKBS;508
10.3.2.5;ANALYZING AND COMPILING THE RTKBS;509
10.3.2.6;CONCLUSIONS;511
10.3.2.7;REFERENCES;511
10.3.3;CHAPTER 78. RIGAS: AN EXPERT SERVER TASK IN REALTIME ENVIRONMENTS;512
10.3.3.1;INTRODUCTION;512
10.3.3.2;ARCHITECTURE;513
10.3.3.3;RIGAS ARCHITECTURE;514
10.3.4;CHAPTER 79. DICE: A REAL-TIME TOOLBOX;518
10.3.4.1;1 Introduction;518
10.3.4.2;2 Real-time issues;518
10.3.4.3;3 The blackboard solution;518
10.3.4.4;4 Knowledge representation;519
10.3.4.5;5. Knowledge compilation and reasoning mechanisms;520
10.3.4.6;6. Implementation of DICE using VAX/VMS ;521
10.3.4.7;7. Example using DICE;521
10.3.4.8;8. Conclusions;521
10.3.4.9;References;521
10.4;Section XIII: Development of Real-time AI Systems;524
10.4.1;CHAPTER 80. LOW-COST ENVIRONMENT FOR ANALYSIS AND DESIGN OF KNOWLEDGE-BASED CONTROL ALGORITHMS;524
10.4.1.1;INTRODUCTION;524
10.4.1.2;SYSTEM RECUIREMENTTS;524
10.4.1.3;REXCON ENVIRONMENT;525
10.4.1.4;CONCLUDING REMARKS;528
10.4.1.5;ACKNOWIEDGEMENTS;528
10.4.1.6;REFERENCES;528
10.4.2;CHAPTER 81. PREDICTING AND IMPROVING RESPONSE-TIMES OF PERFECT-MODELS;530
10.4.2.1;INTRODUCTION;530
10.4.2.2;A PERFECT-MODEL OF THE GENERIC NUCLEAR PLANT;530
10.4.2.3;DIAGNOSING PLANT DISTURBANCES WITH MFM-MODELS IN PERFECT;531
10.4.2.4;FUNCTIONALITY OF THE ANALYZER;532
10.4.2.5;CALCULATING WORTH;533
10.4.2.6;CONCLUSIONS AND FURTHER WORK;535
10.4.2.7;REFERENCES;535
10.4.3;CHAPTER 82. AN EXECUTION ENVIRONMENT FOR REAL-TIME MODEL-BASED SUPERVISORY CONTROL AND DIAGNOSTIC SYSTEMS;536
10.4.3.1;INTRODUCTION;536
10.4.3.2;THE COMPUTATIONAL MODEL;537
10.4.3.3;RUNNING IN REAL-TIME;537
10.4.3.4;THE IMPLEMENTATION;538
10.4.3.5;AN EXAMPLE: ROBOT DIAGNOSTICS;540
10.4.3.6;CONCLUSIONS;541
10.4.3.7;REFERENCES;541
11;AUTHOR INDEX;542
12;KEYWORD INDEX;544