Computationally Intelligent Hybrid Systems: The Fusion of Soft Computing and Hard Computing

This uniquely crafted work combines the experience of many internationally recognized experts in the soft- and hard-computing research worlds to present practicing engineers with the broadest possible array of methodologies for developing innovative and competitive solutions to real-world problems. Each of the chapters illustrates the wide-ranging applicability of the fusion concept in such critical areas as Computer security and data mining Electrical power systems and large-scale plants Motor drives and tool wear monitoring User interfaces and the World Wide Web Aerospace and robust control This must-have guide for practicing engineers, researchers, and R&D managers who wish to create or understand computationally intelligent hybrid systems is also an excellent primary source for graduate courses in soft computing, engineering applications of artificial intelligence, and related topics.

Table of Contents:
Contributors xv Foreword xvii David B. Fogel Preface xix Editor’s Introduction to Chapter 1 1 1 Introduction to Fusion of Soft Computing and Hard Computing 5 Seppo J. Ovaska 1.1 Introduction 5 1.1.1 Soft Computing 5a 1.1.2 Fusion of Soft-Computing and Hard-Computing Methodologies 7 1.2 Structural Categories 9 1.2.1 Soft Computing and Hard Computing Are Isolated from Each Other 10 1.2.2 Soft Computing and Hard Computing Are Connected in Parallel 11 1.2.3 Soft Computing with Hard-Computing Feedback and Hard Computing with Soft-Computing Feedback 12 1.2.4 Soft Computing is Cascaded with Hard Computing or Hard Computing is Cascaded with Soft Computing 12 1.2.5 Soft-Computing-Designed Hard Computing and Hard-Computing-Designed Soft Computing 13 1.2.6 Hard-Computing-Augmented Soft Computing and Soft-Computing-Augmented Hard Computing 14 1.2.7 Hard-Computing-Assisted Soft Computing and Soft-Computing-Assisted Hard Computing 15 1.2.8 Supplementary Categories 16 1.2.9 General Soft-Computing and Hard-Computing Mapping Functions 19 1.3 Characteristic Features 19 1.3.1 Proportional Integral Derivative Controllers 20 1.3.2 Physical Models 20 1.3.3 Optimization Utilizing Local Information 21 1.3.4 General Parameter Adaptation Algorithm 22 1.3.5 Stochastic System Simulators 22 1.3.6 Discussion and Extended Fusion Schemes 22 1.4 Characterization of Hybrid Applications 24 1.5 Conclusions and Discussion 25 References 27 Editor’s Introduction to Chapter 2 31 2 General Model for Large-Scale Plant Application 35 Akimoto Kamiya 2.1 Introduction 35 2.2 Control System Architecture 36 2.3 Forecasting of Market Demand 37 2.4 Scheduling of Processes 39 2.4.1 Problem Decomposition 39 2.4.2 Hybrid Genetic Algorithms 42 2.4.3 Multiobjective Optimization 43 2.5 Supervisory Control 45 2.6 Local Control 47 2.7 General Fusion Model and Fusion Categories 49 2.8 Conclusions 51 References 51 Editor’s Introduction to Chapter 3 57 3 Adaptive Flight Control: Soft Computing with Hard Constraints 61 Richard E. Saeks 3.1 Introduction 61 3.2 The Adaptive Control Algorithms 62 3.2.1 Adaptive Dynamic Programming 63 3.2.2 Neural Adaptive Control 64 3.3 Flight Control 67 3.4 X-43A-LS Autolander 68 3.5 LOFLYTEw Optimal Control 73 3.6 LOFLYTEw Stability Augmentation 76 3.7 Design for Uncertainty with Hard Constraints 82 3.8 Fusion of Soft Computing and Hard Computing 85 3.9 Conclusions 85 References 86 Editor’s Introduction to Chapter 4 89 4 Sensorless Control of Switched Reluctance Motors 93 Adrian David Cheok 4.1 Introduction 93 4.2 Fuzzy Logic Model 95 4.2.1 Measurement of Flux Linkage Characteristics 95 4.2.2 Training and Validation of Fuzzy Model 97 4.3 Accuracy Enhancement Algorithms 101 4.3.1 Soft-Computing-Based Optimal Phase Selection 102 4.3.2 Hard-Computing-Based On-Line Resistance Estimation 104 4.3.3 Polynomial Predictive Filtering 105 4.4 Simulation Algorithm and Results 108 4.5 Hardware and Software Implementation 109 4.5.1 Hardware Configuration 109 4.5.2 Software Implementation 110 4.6 Experimental Results 111 4.6.1 Acceleration from Zero Speed 112 4.6.2 Low-Current Low-Speed Test 113 4.6.3 High-Speed Test 114 4.6.4 Test of Step Change of Load 118 4.7 Fusion of Soft Computing and Hard Computing 119 4.8 Conclusion and Discussion 122 References 122 Editor’s Introduction to Chapter 5 125 5 Estimation of Uncertainty Bounds for Linear and Nonlinear Robust Control 129 Gregory D. Buckner 5.1 Introduction 129 5.2 Robust Control of Active Magnetic Bearings 130 5.2.1 Active Magnetic Bearing Test Rig 132 5.3 Nominal H1 Control of the AMB Test Rig 133 5.3.1 Parametric System Identification 133 5.3.2 Uncertainty Bound Specification 135 5.3.3 Nominal H1 Control: Experimental Results 137 5.4 Estimating Modeling Uncertainty for H1 Control of the AMB Test Rig 138 5.4.1 Model Error Modeling 140 5.4.2 Intelligent Model Error Identification 141 5.4.3 Uncertainty Bound Specification 146 5.4.4 Identified H1 Control: Experimental Results 147 5.5 Nonlinear Robust Control of the AMB Test Rig 148 5.5.1 Nominal Sliding Mode Control of the AMB Test Rig 148 5.5.2 Nominal SMC: Experimental Results 150 5.6 Estimating Model Uncertainty for SMC of the AMB Test Rig 151 5.6.1 Intelligent System Identification 151 5.6.2 Intelligent Model Error Identification 155 5.6.3 Intelligent SMC: Experimental Results 156 5.7 Fusion of Soft Computing and Hard Computing 159 5.8 Conclusion 162 References 162 Editor’s Introduction to Chapter 6 165 6 Indirect On-Line Tool Wear Monitoring 169 Bernhard Sick 6.1 Introduction 169 6.2 Problem Description and Monitoring Architecture 172 6.3 State of the Art 176 6.3.1 Monitoring Techniques Based on Analytical Models 176 6.3.2 Monitoring Techniques Based on Neural Networks 178 6.3.3 Monitoring Techniques Based on Fusion of Physical and Neural Network Models 181 6.4 New Solution 184 6.4.1 Solution Outline 184 6.4.2 Physical Force Model at Digital Preprocessing Level 185 6.4.3 Dynamic Neural Network at Wear Model Level 187 6.5 Experimental Results 189 6.6 Fusion of Soft Computing and Hard Computing 192 6.7 Summary and Conclusions 194 References 195 Editor’s Introduction to Chapter 7 199 7 Predictive Filtering Methods for Power Systems Applications 203 Seppo J. Ovaska 7.1 Introduction 203 7.2 Multiplicative General-Parameter Filtering 205 7.3 Genetic Algorithm for Optimizing Filter Tap Cross-Connections 207 7.4 Design of Multiplierless Basis Filters by Evolutionary Programming 211 7.5 Predictive Filters for Zero-Crossings Detector 213 7.5.1 Single 60-Hz Sinusoid Corrupted by Noise 213 7.5.2 Sequence of 49-, 50-, and 51-Hz Sinusoids Corrupted by Noise 217 7.5.3 Discussion of Zero-Crossings Detection Application 222 7.6 Predictive Filters for Current Reference Generators 223 7.6.1 Sequence of 49-, 50-, and 51-Hz Noisy Sinusoids 225 7.6.2 Sequence of 49-, 50-, and 51-Hz Noisy Sinusoids Corrupted by Harmonics 229 7.6.3 Artificial Current Signal Corrupted by Odd Harmonics 230 7.6.4 Discussion of Current Reference Generation Application 232 7.7 Fusion of Soft Computing and Hard Computing 233 7.8 Conclusion 234 References 237 Appendix 7.1: Coefficients of 50-Hz Sinusoid-Predictive FIR Filters 239 Editor’s Introduction to Chapter 8 241 8 Intrusion Detection for Computer Security 245 Sung-Bae Cho and Sang-Jun Han 8.1 Introduction 245 8.2 Related Works 247 8.2.1 Neural Computing 248 8.2.2 Genetic Computing 249 8.2.3 Fuzzy Logic 251 8.2.4 Probabilistic Reasoning 253 8.3 Intrusion Detection with Hybrid Techniques 253 8.3.1 Overview 254 8.3.2 Preprocessing with Self-Organizing Map 254 8.3.3 Behavior Modeling with Hidden Markov Models 256 8.3.4 Multiple Models Fusion by Fuzzy Logic 259 8.4 Experimental Results 261 8.4.1 Preprocessing 261 8.4.2 Modeling and Intrusion Detection 263 8.5 Fusion of Soft Computing and Hard Computing 267 8.6 Concluding Remarks 268 References 270 Editor’s Introduction to Chapter 9 273 9 Emotion Generating Method On Human–Computer Interfaces 277 Kazuya Mera and Takumi Ichimura 9.1 Introduction 277 9.2 Emotion Generating Calculations Method 279 9.2.1 Favorite Value Database 280 9.2.2 Calculation Pleasure/Displeasure for an Event 282 9.2.3 Favorite Value of Modified Element 284 9.2.4 Experimental Result 285 9.2.5 Complicated Emotion Allocating Method 286 9.2.6 Dependency Among Emotion Groups 294 9.2.7 Example of Complicated Emotion Allocating Method 296 9.2.8 Experimental Results 297 9.3 Emotion-Oriented Interaction Systems 298 9.3.1 Facial Expression Generating Method by Neural Network 298 9.3.2 Assign Rules to the Facial Expressions 301 9.4 Applications of Emotion-Oriented Interaction Systems 302 9.4.1 JavaFaceMail 302 9.4.2 JavaFaceChat 307 9.5 Fusion of Soft Computing and Hard Computing 308 9.6 Conclusion 310 References 311 Editor’s Introduction to Chapter 10 313 10 Introduction to Scientific Data Mining: Direct Kernel Methods and Applications 317 Mark J. Embrechts, Boleslaw Szymanski, and Karsten Sternickel 10.1 Introduction 317 10.2 What is Data Mining? 318 10.2.1 Introduction to Data Mining 318 10.2.2 Scientific Data Mining 320 10.2.3 The Data Mining Process 321 10.2.4 Data Mining Methods and Techniques 322 10.3 Basic Definitions for Data Mining 323 10.3.1 The MetaNeural Data Format 323 10.3.2 The “Standard Data Mining Problem” 326 10.3.3 Predictive Data Mining 329 10.3.4 Metrics for Assessing Model Quality 333 10.4 Introduction to Direct Kernel Methods 335 10.4.1 Data Mining and Machine Learning Dilemmas for Real-World Data 335 10.4.2 Regression Models Based on the Data Kernel 338 10.4.3 Kernel Transformations 339 10.4.4 Dealing with Bias: Centering the Kernel 340 10.5 Direct Kernel Ridge Regression 342 10.5.1 Overview 342 10.5.2 Choosing the Ridge Parameter 343 10.6 Case Study #1: Predicting the Binding Energy for Amino Acids 344 10.7 Case Study #2: Predicting the Region of Origin for Italian Olive Oils 346 10.8 Case Study #3: Predicting Ischemia from Magnetocardiography 350 10.8.1 Introduction to Magnetocardiography 350 10.8.2 Data Acquisition and Preprocessing 351 10.8.3 Predictive Modeling for Binary Classification of Magnetocardiograms 351 10.8.4 Feature Selection 358 10.9 Fusion of Soft Computing and Hard Computing 359 10.10 Conclusions 359 References 360 Editor’s Introduction to Chapter 11 363 11 World Wide Web Usage Mining 367 Ajith Abraham 11.1 Introduction 367 11.2 Daily and Hourly Web Usage Clustering 372 11.2.1 Ant Colony Optimization 372 11.2.2 Fuzzy Clustering Algorithm 374 11.2.3 Self-Organizing Map 376 11.2.4 Analysis of Web Data Clusters 377 11.3 Daily and Hourly Web Usage Analysis 378 11.3.1 Linear Genetic Programming 379 11.3.2 Fuzzy Inference Systems 382 11.3.3 Experimentation Setup, Training, and Performance Evaluation 387 11.4 Fusion of Soft Computing and Hard Computing 389 11.5 Conclusions 393 References 394 Index 397 About the Editor 409