Yu Advances in Computational Intelligence

1. Auflage 2009
ISBN: 978-3-642-03156-4
Verlag: Springer
Format: PDF
Kopierschutz: Adobe DRM (»Systemvoraussetzungen)

E-Book, Englisch, 530 Seiten

This book constitutes the proceedings of the second International Workshop on Advanced Computational Intelligence (IWACI 2009), with a sequel of IWACI 2008 successfully held in Macao, China. IWACI 2009 provided a high-level international forum for scientists, engineers, and educators to present state-of-the-art research in computational intelligence and related fields. Over the past decades, computational intelligence community has witnessed t- mendous efforts and developments in all aspects of theoretical foundations, archit- tures and network organizations, modelling and simulation, empirical study, as well as a wide range of applications across different domains. IWACI 2009 provided a great platform for the community to share their latest research results, discuss critical future research directions, stimulate innovative research ideas, as well as facilitate inter- tional multidisciplinary collaborations. IWACI 2009 received 146 submissions from about 373 authors in 26 countries and regions (Australia, Brazil, Canada, China, Chile, Hong Kong, India, Islamic Republic of Iran, Japan, Jordan, Macao, Malaysia, Mexico, Pakistan, Philippines, Qatar, Republic of Korea, Singapore, South Africa, Sri Lanka, Spain, Taiwan, Thailand, UK, USA, Ve- zuela, Vietnam, and Yemen) across six continents (Asia, Europe, North America, South America, Africa, and Oceania). Based on the rigorous peer reviews by the Program Committee members, 52 high-quality papers were selected for publication in this book, with an acceptance rate of 36.3%. These papers cover major topics of the theoretical research, empirical study, and applications of computational intelligence.

Autoren/Hrsg.

Yu, Wen

Inhaltsverzeichnis

1;Title Page;2
2;Preface;6
3;Organization;7
4;Table of Contents;9
5;Session 1. Neural Networks;14
5.1;Multi Lingual Speaker Recognition Using Artificial Neural Network;14
5.1.1;Introduction;14
5.1.1.1;Back Propagation Training;15
5.1.1.2;Clustering;15
5.1.2;Previous Work;16
5.1.3;Features of Speech;16
5.1.3.1;Cepstrum Coefficient;16
5.1.3.2;Average PSD;16
5.1.3.3;No. of Zero Crossing;16
5.1.3.4;Length of File;17
5.1.4;Approach;17
5.1.5;Results;19
5.1.6;Conclusion;21
5.1.7;References;21
5.2;Modeling Huntington's Disease Considering the Theory of Central Pattern Generators (CPG);23
5.2.1;Introduction;23
5.2.2;References;31
5.3;Prophetia: Artificial Intelligence for TravelBox$^{®}$ Technology;32
5.3.1;Introduction;32
5.3.1.1;Probability Prediction;33
5.3.1.2;Package Recognition;33
5.3.1.3;Customer Interest Prediction;34
5.3.2;Theory;35
5.3.2.1;Neural Networks;35
5.3.2.2;Self Organizing Maps;36
5.3.2.3;Association Rule Mining;36
5.3.3;Implementation;36
5.3.3.1;Probability Prediction;37
5.3.3.2;Package Recognition;38
5.3.3.3;Customer Interest Prediction;39
5.3.4;Results;40
5.3.4.1;Probability Calculation;40
5.3.4.2;Package Recognition;42
5.3.4.3;Customer Interest Prediction;43
5.3.5;Conclusion;44
5.3.6;Confidentiality;44
5.3.7;References;44
5.4;Application Research of Local Support Vector Machines in Condition Trend Prediction of Reactor Coolant Pump;46
5.4.1;Introduction;46
5.4.2;Principle of SVMs Regression;47
5.4.3;Optimization of Hyperparameters;48
5.4.4;Application of SVMs in Predicting Behavior of RCP;50
5.4.5;Conclusions;53
5.4.6;References;53
5.5;Asymptotic Synchronization for Pulse-Coupled Oscillators with Delayed Excitatory Coupling Is Impossible;55
5.5.1;Introduction;55
5.5.2;Model;56
5.5.3;Preliminaries;58
5.5.4;Main Results;59
5.5.5;Conclusions;60
5.5.6;References;61
5.6;Missing Data Imputation Through the Use of the Random Forest Algorithm;62
5.6.1;Introduction;62
5.6.2;Missing Data;63
5.6.2.1;Missing Data: Categorisation and Mechanism;63
5.6.2.2;Dealing with Missing Data;63
5.6.3;Background;63
5.6.3.1;Random Forests;64
5.6.3.2;Other Paradigms;64
5.6.4;Methodology and System Topologies;65
5.6.5;Data Evaluation and Preprocessing;67
5.6.6;Comparison and Results;67
5.6.7;Discussion and Recommendations for Future Work;70
5.6.8;Conclusion;70
5.6.9;References;70
5.7;Ubiquitous Middleware Using Mobility Prediction Based on Neuro-Association Mining for Adaptive Distributed Object System;72
5.7.1;Introduction;72
5.7.2;Related Works;73
5.7.2.1;Ubiquitous Middleware;73
5.7.2.2;Mobility Prediction;74
5.7.3;Ubiquitous Middleware for Adaptive Distributed Object System;74
5.7.4;Dynamic Replication Based on Neuro-Apriori Algorithm;77
5.7.5;Experimental Evaluation;78
5.7.5.1;Generation of the Rule Nodes;78
5.7.5.2;Performance of the Proposed Algorithm;79
5.7.6;Conclusions and Future Work;80
5.7.7;References;80
5.8;A Growing Algorithm for RBF Neural Network;82
5.8.1;Introduction;82
5.8.2;The Sensitivity Analysis (SA) of Model Output for RBFNN;83
5.8.3;The Growing Method for Selecting Hidden Nodes of RBF NN;84
5.8.3.1;Selecting Hidden Nodes;84
5.8.3.2;Parameters Adjusting;87
5.8.3.3;Growing RBF Neural Network;88
5.8.4;Simulations;88
5.8.5;Conclusion;91
5.8.6;References;91
5.9;Fault Tolerance Based on Neural Networks for the Intelligent Distributed Framework;92
5.9.1;Introduction;92
5.9.2;Related Works;93
5.9.2.1;Fault Tolerance in Distributed Objects;93
5.9.2.2;Implementing Failure Detectors;94
5.9.3;Structure of the Intelligent Distributed Framework;94
5.9.3.1;Components of the Fault Tolerant Scheme;95
5.9.4;Selection of Alternative Object Based on Neural Networks;96
5.9.5;Simulation Results;99
5.9.6;Conclusion;100
5.9.7;References;100
5.10;Learning RNN-Based Gene Regulatory Networks for Robot Control;102
5.10.1;Introduction;102
5.10.2;Modeling GRNs for Robot Control;103
5.10.2.1;RNN-Based Regulatory Model;104
5.10.2.2;Learning Algorithm for Constructing GRN Controllers;105
5.10.2.3;Robot Programming by Demonstration;106
5.10.3;Experiments and Results;107
5.10.3.1;Modeling GRNs;107
5.10.3.2;Learning GRNs for Robot Control;108
5.10.4;Conclusions and Future Work;110
5.10.5;References;111
5.11;Fault Detection for Networked Control Systems via Minimum Error Entropy Observer;112
5.11.1;Introduction;112
5.11.2;System Description and Problem Formulation;113
5.11.3;Design of Residual Generator;114
5.11.4;Fault Detection;116
5.11.5;Illustrative Example;116
5.11.6;Conclusions;118
5.11.7;References;119
5.12;Discrete-Time Reduced Order Neural Observers;121
5.12.1;Introduction;121
5.12.2;Preliminaries;123
5.12.2.1;Nonlinear Reduced Order Observers;123
5.12.2.2;Discrete-Time Recurrent High Order Neural Networks;124
5.12.2.3;The EKF Training Algorithm;125
5.12.3;Discrete-Time Reduced Order Neural Observers;126
5.12.3.1;RONO for the Van Der Pol Oscillator;128
5.12.4;Conclusions;128
5.12.5;References;129
5.13;A New Neural Observer for an Anaerobic Wastewater Treatment Process;131
5.13.1;Introduction;131
5.13.1.1;Brief Review of the State of the Art;131
5.13.2;Anaerobic Digestion Process;132
5.13.2.1;Process Description;132
5.13.2.2;Problem Statement;133
5.13.3;Neural Networks;134
5.13.3.1;Discrete-Time Recurrent High Order Neural Network;134
5.13.3.2;The Extended Kalman Filter as Training Algorithm;135
5.13.4;RHONO for Biomass and Substrate Estimation;136
5.13.4.1;Observer Design;136
5.13.4.2;Tuning Guidelines;137
5.13.5;Results and Discussion;137
5.13.6;Conclusions;139
5.13.7;References;139
5.14;Prediction of Protein Subcellular Multi-localization by Using a Min-Max Modular Support Vector Machine;141
5.14.1;Introduction;141
5.14.2;Methods;143
5.14.2.1;Classification ofMulti-label Problems;143
5.14.2.2;Task Decomposition;145
5.14.2.3;Feature Extraction;146
5.14.3;Results and Discussion;146
5.14.4;Conclusion;149
5.14.5;References;150
5.15;Application of MultiLayer Perceptron Type Neural Network to Camera Calibration;152
5.15.1;Introduction;152
5.15.2;Conventional Calibration Method;153
5.15.3;Implicit Camera Calibration;155
5.15.3.1;Calibration Method Using ANN;155
5.15.3.2;MLPNN Structure for Camera Calibration;156
5.15.4;Experimental Results;157
5.15.5;Conclusion;160
5.15.6;References;161
5.16;Hierarchical Neural Network Model for Water Quality Prediction in Wastewater Treatment Plants;162
5.16.1;Introduction;162
5.16.2;Dynamic Model of Wastewater Treatment Process;164
5.16.3;Reaction Rates Identification via Neural Network;166
5.16.4;Soft-Sensor of Water Quality via Hierarchical Neural Networks;167
5.16.5;Application to a Wastewater Treatment Plant;170
5.16.6;Conclusions;172
5.16.7;References;172
5.17;Third Generation Neural Networks: Spiking Neural Networks;174
5.17.1;Introduction;174
5.17.2;Information Encoding and Evolution of Spiking Neurons;175
5.17.3;Mechanism of Spike Generation in Spiking Neurons;176
5.17.4;Models of Spiking Neurons;179
5.17.5;Spiking Neural Networks (SNNs);180
5.17.6;Concluding Remarks;182
5.17.7;References;182
6;Session 2. Fuzzy Systems;186
6.1;Choquet Fuzzy Integral Applied to Stereovision Matching for Fish-Eye Lenses in Forest Analysis;186
6.1.1;Introduction;186
6.1.2;Design of the Matching Process;188
6.1.2.1;Epipolar: System Geometry;188
6.1.2.2;Similarity: Attributes for Area and Feature-Based;189
6.1.2.3;Uniqueness: Applying the Choquet Fuzzy Integral Paradigm;190
6.1.3;Results;191
6.1.4;Concluding Remarks;193
6.1.5;References;194
6.2;Fuzzy OLAP: A Formal Definition;195
6.2.1;Introduction;195
6.2.2;Motivating Example;195
6.2.3;Fuzzy Multidimensional Model;197
6.2.4;Level Climbing of the Fuzzy Cube;201
6.2.5;Conclusion;202
6.2.6;References;203
6.3;Caller Behaviour Classification: A Comparison of SVM and FIS Techniques;205
6.3.1;Introduction;205
6.3.2;The Developed System;207
6.3.3;Selection and Preprocessing of Data;208
6.3.4;Support Vector Machine Field Classifiers;209
6.3.5;Fuzzy Inference System Field Classifiers;210
6.3.6;Comparison of the Support Vector Machine and Fuzzy Inference System Field Classifiers;211
6.3.7;Conclusion;213
6.3.8;References;213
6.4;A Dual-Model Discrete-Time Jumping Fuzzy System Approach to NCS Design;215
6.4.1;Introduction;215
6.4.2;Dual-Mode Discrete-Time Jumping Fuzzy Model;216
6.4.3;Guaranteed Cost Controller Design;217
6.4.4;Simulation Examples;221
6.4.5;Conclusions;223
6.4.6;References;224
6.5;A Continuous-Time Recurrent Neurofuzzy Network for Black-Box Modeling of Insulin Dynamics in Diabetic Type-1 Patients;225
6.5.1;Introduction;225
6.5.2;Problem Statement;226
6.5.3;Proposed Recurrent Neurofuzzy Network and Training Algorithm;228
6.5.3.1;Observation-Training Algorithm;229
6.5.4;Modeling and Identification of Insulin Dynamics in Different Patients;230
6.5.4.1;Simulation Results under Unrestricted Sampling Conditions;232
6.5.4.2;Simulation Results under Restricted Conditions;232
6.5.5;Conclusions;232
6.5.6;References;233
6.6;Vague Query Based on Vague Relational Model;235
6.6.1;Introduction;235
6.6.2;Basic Knowledge;236
6.6.3;Vague Relational Model;238
6.6.4;Vague Data Redundancies and Removal;239
6.6.4.1;Similarity Measure of Vague Data;239
6.6.4.2;Data Redundancies;239
6.6.5;Vague Querying with SQL;240
6.6.6;Conclusions;243
6.6.7;References;243
6.7;Identification and Speed Control of a DC Motor Using an Input-Output Recurrent Neurofuzzy Network;245
6.7.1;Introduction;245
6.7.2;System Identification;246
6.7.3;Input-Output Recurrent Neurofuzzy Network;246
6.7.3.1;Structure;246
6.7.3.2;Linearization of the Antecedent Parameters;247
6.7.3.3;Nonlinear Constraints for the Parameters;248
6.7.3.4;Parameter Initialization Algorithm;249
6.7.3.5;Training Using Kalman Filter;250
6.7.3.6;Training Using Steepest Descent Algorithm;250
6.7.3.7;Certain Equivalence Control;251
6.7.4;Experimental Results;251
6.7.5;Conclusions;253
6.7.6;References;254
6.8;Hybrid Intelligent Control Scheme for an Anaerobic Wastewater Treatment Process;255
6.8.1;Introduction;255
6.8.2;Anaerobic Digestion Process;256
6.8.2.1;Process Description;256
6.8.2.2;Problem Statement;257
6.8.3;Neural Networks Observer for Biomass and Substrate Estimation;258
6.8.3.1;Observer Development;258
6.8.3.2;Validation;260
6.8.4;Hybrid Intelligent Control Scheme;260
6.8.4.1;Design of a Control Strategy;260
6.8.4.2;Validation;261
6.8.5;Conclusions;263
6.8.6;References;263
7;Session 3. Evolutionary Algorithms;265
7.1;Workability of a Genetic Algorithm Driven Sequential Search for Eigenvalues and Eigenvectors of a Hamiltonian with or without Basis Optimization;265
7.1.1;Introduction;265
7.1.2;The Method;266
7.1.2.1;Diagonalization in a Fixed Basis : Lowest Eigenvalue and Vector;266
7.1.2.2;Finding Higher Eigenvalues and Eigenvectors in a Fixed Basis;268
7.1.2.3;Diagonalization with Basis Optimization : Lowest Eigenvalue;268
7.1.3;Results and Discussion;269
7.1.3.1;Ground and Excited Eigenvalues in a Fixed Basis;269
7.1.3.2;Diagonalization of Hamiltonian with Simultaneous Optimization of Basis Parameters;272
7.1.4;Conclusion;273
7.1.5;References;274
7.2;An Improved Quantum Evolutionary Algorithm Based on Artificial Bee Colony Optimization;275
7.2.1;Introduction;275
7.2.2;Main Process of Basic QEA;277
7.2.2.1;Qubit Chromosome;277
7.2.2.2;Quantum Mutation;277
7.2.2.3;Quantum Whole Interference Crossover;278
7.2.3;The Proposed Hybrid QEA Based on ABC;278
7.2.3.1;Artificial Bee Colony(ABC) Optimization;278
7.2.3.2;The Proposed Hybrid QEA Based on ABC;279
7.2.4;Experimental Results;281
7.2.5;Conclusions;283
7.2.6;References;283
7.3;Rough Approximation Operators with Hedges;285
7.3.1;Introduction;285
7.3.2;Preliminaries;286
7.3.2.1;Rough Sets;286
7.3.2.2;L-Sets;286
7.3.2.3;L-Concept Lattice Introduced by R. B\u{e}lohl\'{a}vek;287
7.3.3;Rough Operators with Hedges;288
7.3.4;Generalized Framework;291
7.3.5;Related to Concept Lattice Induced by R. B\u{e}lohl\'{a}vek;292
7.3.6;Conclusion;293
7.3.7;References;293
7.4;An Evolutionary Algorithm with Lower-Dimensional Crossover for Solving Constrained Engineering Optimization Problems;295
7.4.1;Introduction;295
7.4.2;Framework of the New EA (LDNSEA);296
7.4.3;Application in Constrained Optimization;297
7.4.4;Numberical Experimental Result;300
7.4.5;Conclusion;303
7.4.6;References;303
7.5;Gene Regulatory Network Reconstruction of P38 MAPK Pathway Using Ordinary Differential Equation with Linear Regression Analysis;305
7.5.1;Introduction;305
7.5.2;Data;307
7.5.3;Method;308
7.5.3.1;Fitting Selection of Connections;309
7.5.3.2;Procedure of the Algorithm;309
7.5.4;Experiment;309
7.5.4.1;Reconstruction of P38 GRN;310
7.5.4.2;Software;312
7.5.5;Conclusion;312
7.5.6;References;313
7.6;A Novel Multi-threshold Segmentation Approach Based on Artificial Immune System Optimization;315
7.6.1;Introduction;315
7.6.2;Gaussian Approximation;317
7.6.3;Clonal Selection Algorithm;318
7.6.4;Determination of Thresholding Values;319
7.6.5;Implementation Details;320
7.6.6;Conclusions;322
7.6.7;References;322
7.7;Research on Improvement Strategies and ParameterAnalysis of Ant Colony Algorithm for One-Dimensional Cutting Stock Problem;324
7.7.1;Introduction;324
7.7.2;Improvement Strategies of Ant Colony Algorithm;325
7.7.2.1;Efficiency Improvement Strategies;325
7.7.2.2;Solving Ability Improvement Strategies;326
7.7.3;Description of One-Dimensional Cutting Stock Problem;328
7.7.4;Improved Ant Colony Algorithm for One-Dimension Cutting Stock Problem (IACA-1CSP);329
7.7.4.1;Parts Encoding;329
7.7.4.2;Solution Path;329
7.7.4.3;Specific Implementation Steps;329
7.7.5;Parameter Analysis;330
7.7.5.1;Number of Ants $m$;331
7.7.5.2;Pheromone Intensity $Q$;331
7.7.5.3;Pheromone Heuristic Factor $\alpha$;331
7.7.5.4;Expectation Heuristic Factor $\beta$;332
7.7.5.5;Volatile Factor $\rho$;332
7.7.6;Experimental Results;333
7.7.7;Conclusion;333
7.7.8;References;334
7.8;Mixture of Experts with Genetic Algorithms;335
7.8.1;Introduction;335
7.8.2;Genetic Algorithms;336
7.8.3;Mixture of Experts;336
7.8.3.1;Mixture of Experts with 1-NN;337
7.8.3.2;Mixture of Experts with MNN;337
7.8.4;Experimental Results;338
7.8.5;Concluding Remarks;340
7.8.6;References;341
7.9;Opposition-Based Particle Swarm Optimization with Velocity Clamping (OVCPSO);343
7.9.1;Introduction;343
7.9.2;Related Work;345
7.9.2.1;Essence of Opposition-Based Learning and PSO;345
7.9.2.2;Velocity Clamping in PSO;345
7.9.2.3;Inertia Weight;346
7.9.3;OVCPSO Algorithm;347
7.9.4;Experiments and Results;348
7.9.4.1;Benchmark Functions with Brief Description;348
7.9.4.2;OVCPSO Parameters Initialization;348
7.9.5;Performance Comparisons and Discussion;349
7.9.5.1;OVCPSO and Probability Ranges;351
7.9.6;Conclusion and Future Work;351
7.9.7;References;352
7.10;Designing a Compact Genetic Algorithm with Minimal FPGA Resources;353
7.10.1;Introduction;353
7.10.2;The Compact Genetic Algorithm;354
7.10.3;VHDL Design;354
7.10.4;Finite-State Machine;357
7.10.5;Experiments and Results;359
7.10.6;Conclusions and Future Work;360
7.10.7;References;361
8;Session 4. Intelligent Techniques and Applications;362
8.1;Application of DNA Self-assembly on Maximum Clique Problem;362
8.1.1;Introduction;362
8.1.2;DNA Self-assembly;363
8.1.2.1;DNA Tile;363
8.1.2.2;Molecular Self-assembly Processes;364
8.1.2.3;Programming Self-assembly of DNA Tiling;364
8.1.3;Maximum Clique Problem;364
8.1.4;DNA Self-assembly for MCP;365
8.1.4.1;Non-deterministic Search Cliques;365
8.1.4.2;The Non-deterministic Algorithm for MCP;367
8.1.4.3;Complexity Analysis;368
8.1.5;Conclusions;370
8.1.6;References;370
8.2;Modeling of the Relative Humidity and Control of the Temperature for a Bird Incubator;372
8.2.1;Nomenclature;372
8.2.2;Introduction;373
8.2.3;The Bird Incubator System;373
8.2.4;Mathematical Model of the Temperature of the Incubator;374
8.2.5;Modeling of Relative Humidity Behavior Using Functional Networks;375
8.2.6;Simulations;377
8.2.7;Conclusions;379
8.2.8;References;379
8.3;A Novel Automatic Method on Diagnosing Movement Disorders;381
8.3.1;Introduction;381
8.3.2;Materials and Methods;382
8.3.3;Discussions;385
8.3.4;Future Considerations;386
8.3.5;References;386
8.4;The Fault Diagnosis of Electric Railway Traction Substation with Model-Based Diagnosis of Integration of FDI and DX Approaches;388
8.4.1;Introduction;388
8.4.2;The Basic Theory of Model-Based Diagnosis;389
8.4.3;The Diagnosis Method of Traction Substation;390
8.4.4;Traction Substation Modeling;391
8.4.5;Experiment Results and Analysis;392
8.4.6;Discussion;395
8.4.7;Conclusion;395
8.4.8;References;395
8.5;A Single-Hop Active Clustering Algorithm for Wireless Sensor Networks;397
8.5.1;Introduction;397
8.5.2;Related Works;398
8.5.3;SHAC Routing Algorithm;399
8.5.3.1;Network Model;400
8.5.3.2;SHAC Algorithm;401
8.5.3.3;Selecting Tentative Clusterhead;402
8.5.3.4;Active Selecting Clusterheads;402
8.5.3.5;Balancing Cluster Member Energy;403
8.5.4;Simulations and Analysis;404
8.5.5;Conclusions and Future Works;405
8.5.6;References;406
8.6;A Transelevator Moving Inside of an Automatic Warehouse in Virtual Reality;407
8.6.1;Introduction;407
8.6.2;Building of the Physical Model of the Department Store;408
8.6.3;Building of the Logical Model;408
8.6.3.1;Case1: Putting the Load and Taking from One Inferior Place;409
8.6.3.2;Case 2: Putting from the Point of Angle E. Taking from the Point of Angle A;410
8.6.3.3;Case 3: Putting and Taking in Elevation from One Point;410
8.6.3.4;Case 4: Putting and Taking Considering Movements in X Direction, E=A;410
8.6.3.5;Case 5: Putting from a Point of Angle E. Taking the Load in Elevation in de Y Direction;410
8.6.3.6;Case 6: Putting the Load in Elevation in Y Direction. Taking the Load from One Point of Angle A;410
8.6.4;Example of a Case;410
8.6.5;Simulation;411
8.6.6;Conclusion;414
8.6.7;References;414
8.7;Improved AFT and Background-Mesh Generation for FEM with Applications in Microwave;415
8.7.1;Introduction;415
8.7.2;Method;416
8.7.2.1;Two Dimensional Mesh Generation Procedure;416
8.7.2.2;Three Dimensional Mesh Generation Procedure;418
8.7.3;Examples;419
8.7.3.1;Examples of Two Dimensional Meshes Generated;419
8.7.3.2;RCS of Sphere Calculated with FEM;419
8.7.3.3;RCS of Cylinder Calculated with FEM;420
8.7.4;Conclusion;421
8.7.5;References;421
8.8;Application of Non-redundant Association Rules in University Library;422
8.8.1;Introduction;422
8.8.2;Improved Methods of Pruning Redundant Rules;423
8.8.2.1;Related Concepts;423
8.8.2.2;The Definition of Redundancy Rules;424
8.8.2.3;Theorems and Prove of Redundancy in Negative Association Rules;424
8.8.2.4;The Related Concepts of Correlation;425
8.8.3;The Description and Analysis of the Algorithm;426
8.8.4;The Experimental Results;428
8.8.5;Conclusion;429
8.8.6;References;429
8.9;Global Structure Constraint: A Fast Pre-location for Active Appearance Model;431
8.9.1;Introduction;431
8.9.2;Global Structure Constraint Model;432
8.9.2.1;Patches and Landmarks;432
8.9.2.2;Shape Model and Transformation Parameters;432
8.9.2.3;Color Information Model;433
8.9.2.4;Search and Measure;434
8.9.3;Working Together with ASM and AAM;434
8.9.4;Experiment Results;437
8.9.5;Discussion and Conclusion;438
8.9.6;References;438
8.10;Bio-inspired Architecture for Visual Recognition of Humans Walking;440
8.10.1;Introduction;440
8.10.2;Related Works;441
8.10.3;Biological Foundations;442
8.10.4;Architecture;442
8.10.4.1;First Stage (A);443
8.10.4.2;Second Stage (B);444
8.10.4.3;Third Stage (C);445
8.10.5;Results;445
8.10.6;Conclusions and Future Work;447
8.10.7;References;448
8.11;Computational Model for Electric Fault Diagnosis in Induction Motors;450
8.11.1;Introduction;450
8.11.2;Computational Model;450
8.11.2.1;Training Phase;450
8.11.2.2;Execution Phase;451
8.11.3;Mathematical Model of Induction Motor;452
8.11.3.1;Novel Mathematical Model;452
8.11.3.2;Calculation of Varying Parameters for Mathematical Model Proposed;453
8.11.3.3;Fault Injection;454
8.11.4;Methodology to Transform Temporal Response in Patterns;455
8.11.4.1;Pattern Extraction;455
8.11.5;Training and Execution of Artificial Neural Network;457
8.11.5.1;Training of Neural Network;457
8.11.5.2;Evaluation of Neural Network;458
8.11.6;Conclusions;458
8.11.7;References;459
8.12;Closed-Loop Identification of a Nonlinear Servomechanism: Theory and Experiments;460
8.12.1;Introduction;460
8.12.2;Closed-Loop Parameter Identification;461
8.12.2.1;Preliminaires;461
8.12.2.2;Stability Analysis of the Closed-Loop System;462
8.12.2.3;Stability Analysis of the Error Dynamics;463
8.12.2.4;Parameter Convergence;464
8.12.3;Experimental Results;464
8.12.3.1;Parameter Identification without Adding a Disturbance;465
8.12.3.2;Trajectory Tracking Experiments;466
8.12.3.3;Parameter Identification under Constant Disturbances;467
8.12.4;Conclusion;468
8.12.5;References;468
8.13;Dynamic Classifier Selection with Confidence Intervals;469
8.13.1;Introduction;469
8.13.2;Confidence Measures;470
8.13.3;Classifier Selection;471
8.13.3.1;Static Classifier Selection (SCS);471
8.13.3.2;Dynamic Classifier Selection (DCS);471
8.13.4;Dynamic Classifier Selection with Confidences (DCS-CONFI);472
8.13.5;Resampling Methods;473
8.13.5.1;Bagging;473
8.13.5.2;Boosting and Arc-x4;474
8.13.6;Genetic Algorithm;475
8.13.7;Experimental Results;475
8.13.8;Concluding Remarks;476
8.13.9;References;477
8.14;Optimal Neuron-Controller for Fluid Triple-Tank System via Improved ADDHP Algorithm;479
8.14.1;Introduction;479
8.14.2;Standard ADDHP;481
8.14.3;The Improved ADDHP Algorithm;482
8.14.3.1;Training for Critic Network;483
8.14.3.2;Training for Action Network;484
8.14.4;Simulation and Results;485
8.14.4.1;The Plant;485
8.14.4.2;The Design for the Neuron-Controller;485
8.14.4.3;Results of Simulation;486
8.14.5;Conclusion;487
8.14.6;References;488
8.15;Method of Learning for Life Pet Artificial;489
8.15.1;Introduction;489
8.15.1.1;Cathexis;490
8.15.1.2;System Behavior;491
8.15.2;Learning Module;491
8.15.2.1;Description of the Classes That Make Up the Learning Modules;492
8.15.2.2;Sequence Diagram Learning;493
8.15.3;Learning;493
8.15.4;Results;495
8.15.5;Conclusions;496
8.15.6;References;497
8.16;An Sliding Mode Control for an Elbow Arm;499
8.16.1;Introduction;499
8.16.2;Preliminaries;500
8.16.3;Sliding Mode Control with Gravity Compensator;501
8.16.4;Simulation Results;502
8.16.5;Conclusion;503
8.16.6;References;503
8.17;Stabilization on a Physical Pendulum with Moving Mass;505
8.17.1;Introduction;505
8.17.2;Physical Pendulum with Moving Mass;506
8.17.2.1;Lagrangian Modeling;506
8.17.2.2;Model Properties;507
8.17.3;The Control Law;508
8.17.4;Numerical Simulations;511
8.17.4.1;Simulation Analysis;513
8.17.5;Conclusions;513
8.17.6;References;513
8.18;Restricted Growth String for Video-Type Classification;515
8.18.1;Introduction;515
8.18.2;Complexity of Video-Type Clustering;516
8.18.3;Proposed Methodology;517
8.18.3.1;RGS;517
8.18.3.2;Data Clustering;518
8.18.4;Results;520
8.18.5;Conclusion and Future Work;521
8.18.6;References;521
9;Author Index;523