Advances in Neural Network Research and Applications (eBook)

Title Page 2
Preface 5
Organization 7
Table of Contents 16
Prediction and Forecasting 16
A Novel Prediction Mechanism with Modified Data Mining Technique for Call Admission Control in Wireless Cellular Network 26
Introductions 26
Related Works 27
CAC Overviews 27
Time Series Prediction 28
Methodology 28
The Proposed Prediction Method 28
Design Considerations 31
Experimental Results 31
Conclusions 33
References 33
The Study of Forecasting Model of Rock Burst for Acoustic Emission Based on BP Neural Network and Catastrophe Theory 35
Introduction 35
The Establishment of Forecasting Model of Time Series Neural Network 36
The Introduction of BP Neural Network 36
Process of Network Training 36
Establishment of Time Series Neural NetworkForecasting Model 37
Catastrophe Cusp Model and Yield Mechanism 38
Application of BP Neural Network – Mutation Model in Rock Burst Prediction of Acoustic Emission 40
Conclusion 42
References 42
Application of Wavelet Neural Network to Prediction of Water Content in Crude Oil 44
Introduction 44
Water Content in Crude Oil Measurement System 45
Wavelet Neural Network Model 46
Basic Theory 46
Wavelet Neural Network 46
The Prediction Model Simulation 48
Conclusion 50
References 50
Prediction of Urban Heat Island Intensity in Chuxiong City with Backpropagation Neural Network 51
Introduction 51
A Survey of Chuxiong City 52
Natural Situation of Chuxiong City 52
Sequential Variation of Urban Heat Island in Chuxiong City 52
Prediction of Urban Heat Island Intensity in Chuxiong City with Backpropagation Neural Network 53
Influencing Factors of Urban Heat Island in Chuxiong City 53
Prediction Process with Backpropagation Neural Network 53
Predicted Result and Analysis 56
Conclusion 56
References 56
Research on the Fouling Prediction of Heat Exchanger Based on Wavelet Relevance Vector Machine 59
Introduction 59
The Principles of Wavelet Relevance Vector Machine 60
Wavelet Transform 60
Relevance Vector Machine 61
Construction of WRVM Prediction Model 62
Fouling Thermal Resistance Prediction Based on WRVM 63
Data Acquisition and Pretreatment 63
Structure of Kernel Function 64
Results and Discussion 64
Conclusions 67
References 67
Electricity Price Forecasting Using Neural Networks Trained with Entropic Criterion 68
Introduction 68
MLP Trained with MEE Criterion 69
Price Forecasting by MLP with MEE Criterion 70
Simulation and Analysis 72
Conclusions 74
References 74
Least Square Support Vector Machine Ensemble for Daily Rainfall Forecasting Based on Linear and Nonlinear Regression 75
Introduction 75
The Building Process of the LS-SVM Ensemble Model 76
Feature Extraction Used by PP–PSO 76
Extraction of Linear Features by Three Traditional Linear Regression 77
Extraction of Nonlinear Features by Three NNs Methods 78
Nonlinear Ensemble Based on LS-SVM 78
Experiments Analysis 79
Empirical Data 80
The Performance Evaluation of the Model 81
Analysis of the Results 81
Conclusions 83
References 83
Estimating Portfolio Risk Using GARCH-EVT-Copula Model: An Empirical Study on Exchange Rate Market 85
Introduction 85
GARCH-EVT Model and Copula Parameter Estimates 86
GARCH-EVT Model 86
Copula Parameter Estimates 87
Simulation Algorithm and Portfolio Risk Analysis 88
Simulation Algorithm 88
Portfolio Risk Analysis 89
Empirical Studies in Foreign Exchange Market 90
Summary Statistics 90
GARCH-EVT-Copula Application 90
Conclusions 91
References 92
Forecasting Financial Time Series via an Efficient CMAC Neural Network 93
Introduction 93
Literature Review 94
CMAC NN 94
SVR 95
Method 96
Experimental Results 97
Conclusion 100
References 100
Forecasting Daily Cash Turnover of Bank with EWMA and SVR 103
Introduction 103
Forecasting Method with EWMA and SVR 104
Domain Description 104
Exponential Weighted Moving Average 104
Support Vector Regression Algorithm 105
The Forecasting Flow 106
Experiment Results 107
Experimental Data and Performance Measures 107
Primary Results 108
Conclusion and Future Work 109
References 109
Financial Distress Prediction Model via GreyART Network and Grey Model 111
Introduction 111
Preliminaries 112
Grey Relational Analysis 112
GreyART Network 113
Grey Model 114
Financial Distress Prediction Model 114
Data Collection and Variable Selection 114
Data Preprocess 115
Learning and Testing Phases 116
Growing Extraction Method 117
Results and Discussion 118
Conclusions 119
References 120
Risk Assessment Model Based on Immune Theory 121
Introduction 121
Design of Immune Network 121
Initialization of Network and Antigen Presentation 122
Clone and Mutation 122
Termination of Network Training 124
Definition of Risk Measure 125
Case Study 125
Conclusions 127
References 128
Short-Term Load Forecasting Based on Bayes and RS 129
Introduction 129
Bayes Decision Theory 129
Ascertains a Priori Probability Distribution 129
Ascertains a Likelihood Function 130
Calculates Posterior Probability 131
Chooses Input Vector ${/bar /omega}_{j}$ 131
Rough Sets Reduction Algorithm 131
Use the Dicernibility Matrix to Get the Least Reduction 131
Reduction Algorithm Based on the Attribute Dependability 132
Reduction Algorithm Based on Conditional Information Entropy 132
Attribute Reduction Based on Genetic Algorithm 133
Case Analysis 133
Conclusion 135
References 136
Fuzzy Neural Networks 17
Nonlinear System Modeling with a New Fuzzy Model and Neural Compensation 137
Introduction 137
Generating Fuzzy Systems via Kernel Smoothing 139
Neural Compensation for Fuzzy Models 142
Simulation Comparisons 144
Conclusion 146
References 146
A Research of Fuzzy Neural Network in Ferromagnetic Target Recognition 148
Introduction 148
Magnetic Detection System Establish 149
Fuzzy Neural Network 150
Magnetic Field Analysis of the Ferromagnetic Target 150
Target Magnetic Field Model Building 151
Target Characteristic Quantity Extraction 152
FNN Recognition System Design 152
Fuzzy Rule Base Construction 152
Application Example of Ferromagnetic Target Recognition 152
Realization Process of FNN 153
Conclusion 154
References 155
Multiple T-S Fuzzy Neural Networks Soft Sensing Modeling of Flotation Process Based on Fuzzy C-Means Clustering Algorithm 156
Introduction 156
FCM-Based Multiple T-S Fuzzy Neural Network Model 157
FCM Clustering Algorithm 157
Flotation Foam Image Texture Feature Selection and Extraction 159
Dimensionality Reduction Based on PCA 160
T-S Fuzzy Neural Network Model 161
Industrial Experiments 162
Conclusions 163
References 163
A Using Reliability Evaluation Model for Diesel Engine Based on Fuzzy Neural Network 164
Introduction 164
Using Reliability Indices Analysis for Diesel Engine 165
The Using Reliability Evaluation Method Based on Fuzzy Logic 167
BP-Neural-Network-Driven Fuzzy Evaluation Model 169
Case Study 169
Conclusions 170
References 170
Fuzzy Sliding Mode Control with Perturbation Estimation for a Piezoactuated Micromanipulator 172
Introduction 172
System Description 173
Dynamic Modeling and Identification 174
Conventional SMCPE Design 175
Fuzzy SMCPE Design 176
Experimental Studies 177
Conclusions 179
References 179
A Credit Risk Rating Model Based on Fuzzy Neural Network 180
Introduction 180
Literature Review 181
Methodology 182
Data Analysis 183
Acquisition and Preprocessing of Sample Data 183
Construct the Financial Risk Model Based on Back-Propagation Neural Network 183
Construct the Company Scale Rating Based on Fuzzy Inference System 184
Application of Fuzzy Neural Network in Corporate Credit Risk Rating 186
Data Analysis and Assessment 188
Conclusions 189
References 190
Interval-Valued Fuzzy Control 191
Introduction 191
Some Notions 192
Interpolation Mechanism of Interval-Valued Fuzzy Control 193
Simulation Experiment of Interval-Valued Fuzzy Control 195
Conclusion 199
References 200
Research on Fuzzy Preference Relations-Based MAS for Decision Method 202
Introduction 202
Fuzzy Agent 203
The Model of Multi-agent Decision-Making 204
Simulation and Discussion 207
Conclusions 208
References 209
Optimization and Planning 17
Study of Stochastic Demand Inventory Routing Problem with Soft Time Windows Based on MDP 210
Introduction 210
Problem Definition 211
Problem Hypothesis 211
Problem Description 212
MDP Model 212
Solve the Model 214
Example Study 215
Conclusion 216
References 217
An Agent-Based Approach to Joint Procurement Modeling with Virtual Organization 218
Introduction 218
Literature Review 219
Virtual Organization 219
Auction Mechanism 220
Back-Propagation Network (BPN) 220
Methodology and Steps 221
Validation and Results 225
Conclusions 226
References 226
Interactive Hybrid Evolutionary Computation for MEMS Design Synthesis 228
Introduction 228
Interactive Hybrid Computation (IHC) Process 229
Evaluation of the IHC Process 230
Experimental Design 230
Results and Discussion 232
Conclusion and Future Research 234
References 234
Genetic Algorithms for Traffic Grooming in Optical Tree Networks 236
Introduction 236
Problem Definition 237
Upper Bound 237
Algorithms 239
Framework of GAs 239
Crossover and Mutation Operators 239
Approaches of Chromosome Decoding and Fitness Assignment 240
Simulation Results and Analysis 241
Conclusion 243
References 243
Multi-sensor Multi-target Tracking with OOSM 244
Introduction 244
Out-of-Sequence Measurement Update 244
Optimal Update with OOSM 245
Performance Evaluation 247
Conclusion 249
References 250
Hopfield Neural Network Guided Evolutionary Algorithm for Aircraft Penetration Path Planning 251
Introduction 251
HNN for Path Planning 252
Modeling of Penetration Factors 253
Terrain Mask Analysis 253
Terrain Following Analysis 253
Threat Modeling 253
Other Criteria and Restrictions 254
HNN Guided EA 254
EA Based Path Planning Algorithm 254
HNN Guided EA Algorithm 256
Overall Planning Algorithm 256
Simulation Results 257
Conclusion 258
References 259
Fuzzy Material Procurement Planning with Value-at-Risk 260
Introduction 260
Fuzzy MPP Model with VaR 261
Hybrid Algorithm and Numerical Example 263
Hybrid Algorithm 263
Numerical Example 265
Conclusions 265
References 266
Radial Basis Function Network for Endpoint Detection in Plasma Etch Process 268
Introduction 268
Radial Basis Function Network 271
Methodology 273
Simulation and Result 276
Conclusion 277
References 278
A Novel Cellular Neural Network and Its Applications in Motion Planning 279
Introduction 279
The SP-CNN Model 280
Neural Dynamics 280
Motion Planning through SP-CNN 282
Simulations 284
U-Shape Environment 284
Employing Heuristic Information V.S. No Heuristic Information 285
Labyrinth Environment 286
Conclusion 286
References 287
Evaluation of Enterprise ERP System Based on Neural Network Optimized by Ant Colony 288
Introduction 288
Evaluation Indicators System 289
The Concept of ERP 289
Determine the Evaluation Indicators System of the ERP System 289
Ant Colony Neural Network 289
Ant Colony Algorithm 289
Ant Colony Neural Network Model 291
Empirical Analysis 292
Evaluation Criteria of Enterprise ERP System 292
Preprocessing of Raw Data 292
Design of Ant Colony Neural Network and Analysis of Experimental Results 293
Conclusion 295
References 295
Pattern Recognition 18
Visual Attention-Based Ship Detection in SAR Images 296
Introduction 296
Visual Attention-Based Ship Detection 297
Preprocessing 298
The Attention Model 298
Ship Detection 299
Algorithm Summary 299
Experimental Results 300
Conclusion 303
References 304
Recognizing Multi-ships Based on Silhouette in Infrared Image 306
Introduction 306
Drawing Ships Out of Infrared Image 307
Recogniztion of Infrared Ships 308
Analyzing Features of an Infrared Ship 308
The Recognizing Rule Based on BP Neural Network 309
The Algorithm of Recognizing Infrared Ship 310
Simulation Results and Discussion 310
Conclusion 312
References 312
SAR Images Feature Extraction and Recognition Based on G2DCDA 313
Introduction 313
2-Dimensional Clustering-Based Discriminant Analysis (2DCDA) 314
Generalized 2DCDA (G2DCDA) 314
Computational of {/it L} 315
Computational of {/it R} 316
Classification 317
Classification Based 2DCDA Features 317
Classification Based G2DCDA Features 317
Experimental Results 318
Conclusions 320
References 321
Bridge Detection and Recognition in Remote Sensing SAR Images Using Pulse Coupled Neural Networks 322
Introduction 322
Theory and Improvement of PCNN 323
The Basic PCNN Model 323
DLPFPCNN Model Structure 324
Water Area Segmentation Using DLPFPCNN 325
The Sequential Processing of Bridge Recognition 326
Determining the Optimal Results Using Minimum Class Variance Criterion 326
Morphological Clustering 327
Rough Detection of Bridge Outline 328
Conclusions 331
References 331
Approaches to Robotic Vision Control Using Image Pointing Recognition Techniques 332
Introduction 332
Object Tracking and Feature Extraction 333
Extraction of Hand Image 334
Image Processing and Wrist to Remove 334
Finger Search 334
Epipolar Constraint Model 335
Support Vector Machine Classifier 335
Experiment and Valuation 336
Conclusions 338
References 338
A Hybrid Evolutionary Approach to Band Selection for Hyperspectral Image Classification 340
Introduction 340
Initial Band Grouping Based on Conditional Mutual Information 341
Search Algorithm Based on the Combination of SVM and GA 342
Band Pruning with the Branch and Bound Algorithm 343
Experimental Results and Analysis 344
Experimental Results on Washington DC Data Set 344
Experimental Results on Indian Pine Data Set 345
Concluding Remarks 345
References 346
Impulsive Environment Sound Detection by Neural Classification of Spectrogram and Mel-Frequency Coefficient Images 348
Introduction 348
Feature Extraction 349
Spectrogram (log Power Spectrum) 350
Mel Frequency Cepstral Coefficients 351
Data Collection 352
Signal Classification 353
Experimental Results 353
Conclusions 356
References 356
Classification and Control of Cognitive Radios Using Hierarchical Neural Network 358
Introduction 358
Feature Extraction 360
The Hierarchical Neural Network (HNN) 360
Case Study: Protecting the Communication Band of a Legacy Radio 362
Conclusion 363
References 364
Identifying Spatial Patterns of Land Use and Cover Change at Different Scales Based on Self-Organizing Map 365
Introduction 365
Methodology 366
Self-Organizing Map 366
The Procedure of Identifying Spatial Pattern of LUCC Based on SOM 367
Case Study 367
Study Area and Data Pre-processing 367
The Cluster Result at Coarse Scale 369
The Cluster Result at Fine Scale 369
Conclusions 371
References 371
Colon Tumor Microarray Classification Using Neural Network with Feature Selection and Rule-Based Classification 372
Introduction 372
Neural Network with Feature Selection and Rule-Based Classification 373
Parameter Tuning Algorithms 375
Weight Update 375
Membership Function Parameter Update 376
Rule Extraction 376
Experimental Methods and Results 377
Classification by Direct Calculation Using Selected Features 377
Classification by Logical Rule Using Selected Features 378
Comparison with Other Methods 379
Conclusion 380
References 380
Signal and Image Processing 19
Dual Channel Speech Denoising Based on Sparse Representation 382
Introduction 382
Noisy Speech in TF Domain 383
Dual Channel Speech Denoising 383
Experiment and Results 386
Conclusion 387
References 387
Frequency-Domain Blind Separation of Convolutive Speech Mixtures with Energy Correlation-Based Permutation Correction 389
Introduction 389
Typical Complex-Valued ICA Algorithms 390
Kurtosis Maximization 390
c-FastICA 391
Negentropy Maximization 391
Energy Correlation Method for Solving Permutation Problem 391
Speech Characteristic Analysis of Energy Correlation 391
Permutation Solution Based on Energy Correlation 393
Experiments and Results 393
Experiments with Synthetic Speech Signals 394
Experiments with Recorded Speech Signals 396
Conclusion 397
References 398
A Blind Broadband Beamforming Method for Speech Enhancement 399
Introduction 399
Signal Model 399
Blind Broadband Beamforming Method 400
Mean Frame Skewness for Permunication 401
Leaky LMS for Noise Cancellation 401
Experimental Results 402
Conclusions 404
References 404
Algorithm and Simulation Research for Blind Nonlinear System Identification 405
Introduction 405
Hammerstein-Wiener Model Structure 406
Blind Identification Algorithm 407
The Statistical Characteristics of Cyclostationary Signals 407
Identification Steps 407
Simulation Process 409
Conclusion 411
References 411
Study on Digital Image Correlation Using Artificial Neural Networks for Subpixel Displacement Measurement 413
Introduction 413
Description of the Method 414
Verification by Numerical Experiment 416
Generation of Simulated Images 416
Accuracy Evaluation 416
Effect of Speckle Size 418
Effect of Noise 418
Conclusions 419
References 420
Tree Modeling through Range Image Segmentation and 3D Shape Analysis 421
Introduction 421
Related Work 422
Our Approach 422
Segmentation of Leaves and Branches 423
Curvature Estimation 423
Segmentation of Points from Branches and Those from Leaves 423
Segmentation of Points from Different Branches 424
Skeleton Extraction 425
Production of Main Skeleton 425
Connection of Skeleton Nodes 425
Tree Growing 426
Radius Estimation and Generation of Leaves 428
Results 428
Conclusion and Future Work 429
References 429
Combining Bag of Words Model and Information Theoretic Method for Image Clustering 431
Introduction 431
Represent Images with “Codewords” Histogram Generated from BoW 432
Clustering Images Based on IB and Bregman Divergence 432
The Information Bottleneck Principle 433
The Bregman Divergence Clustering Algorithm 434
Experiments and Results 434
Related Work and Discussion 437
References 438
Fast Fuzzy c-Means Clustering Algorithm with Spatial Constraints for Image Segmentation 439
Introduction 439
Fuzzy Clustering with Spatial Constraints(FCM-S) and Its Variants 440
Fast Fuzzy c-Means Clustering Algorithm with Spatial Constraints (FFCM-S) 442
Experimental Results and Analysis 443
Conclusion 445
References 445
Application of Support Vector Machines to Airborne Hyper-Spectral Image Classification 447
Introduction 447
Methodology 448
Principal Components Analysis 448
Support Vector Machines (SVM) 448
Study Site 449
The Jou-Jou Mountain Airborne Hyper-Spectral Image 449
Airborne Image of the Indian Pine Test Site 449
Results and Discussion 449
Purdue's Indian Test Site 449
Jou-Jou Mountain Test Site 450
Conclusions and Recommendations 451
References 452
Research for the Identification Method of the Image Definition Based on a W-N Model 453
Introduction 453
Image Feature Extraction 454
Image Definition Identification 456
Simulation and Experimental Results 457
Conclusion 459
References 459
Analysis of Texture Images Generated by Olfactory System Bionic Model 460
Introduction 460
GMCL and Features 461
GLCM 461
Feature Selection 461
Texture Image Analysis 463
Texture Images Feature Extraction Methods 463
Texture Images Analysis 463
Conclusion 465
References 465
Do Neural Networks Have True Power for Natural Language Processing? 467
Introduction 467
Part of Speech Tagging 467
Introduction 467
POS Tagging Problem 468
POS Tagging with a Hybrid Tagger 468
Experimental Results 471
Self-organizing Documentary Maps for Information Retrieval 472
Introduction 472
Self-organizing Documentary Maps 472
Experimental Results 474
Conclusions 476
References 476
Robust Channel Identification Using FOCUSS Method 477
Introduction 477
Problem Statement 478
Cross Relation (CR) Based SIMO Blind Identification 478
SIMO Identification with Order Overestimation 480
Numerical Experiments and Result Analysis 481
Conclusions 482
References 483
Human Head Modeling Using NURBS Method 484
Introduction 484
Method 485
Image Segmentation 485
Boundary Extraction 485
NURBS Based Surface Reconstruction 486
Result 487
Discussion and Conclusion 488
References 488
Risk Sensitive Unscented Particle Filter for Bearing and Frequency Tracking 490
Introduction 490
Risk Sensitive Filters 491
Risk Sensitive Estimation 491
Risk Sensitive Filter 492
Risk Sensitive Unscented Particle Filter 493
Simulations 494
Conclusion 496
References 496
Fully Complex Multiplicative Neural Network Model and Its Application to Channel Equalization 497
Introduction 497
Channel Equalization 498
Learning Rule for MNN 498
Experiments and Results 502
Conclusions 504
References 504
Robotics and Control 20
Visual Navigation of a Novel Economical Embedded Multi-mode Intelligent Control System for Powered Wheelchair 506
Introduction 506
Structure of Visual Navigation Module 507
Image Data Acquisition 508
Data Processing 509
Adaptive Threshold Algorithm for Marked Line Recognition 509
Abstraction of Marked-Line Center and Removing Noise 509
Feature Parameters Abstraction of Marked-Line 510
Control Strategy 511
Visual Navigation Control Strategy Based on Fuzzy Logic 511
Multi-mode Fusion and Security Measures Related with VM 512
Summary 513
References 513
Neural Networks $L_{2}$-Gain Control for Robot System 515
Introduction 515
The Dynamics of the Robot System and Problem Formulation 516
Function Approximation by Neural Networks 517
Robust Tracking Control and Neural Networks Online Adaptive Adjustment Technique 517
Simulation Study 519
Conclusion 522
References 522
Neural Network Control of Spacecraft Formation Using RISE Feedback 523
Introduction 523
Relative Translational Model 524
Rise Based NN Dynamic Controller 524
Multilayer NN Estimation 524
RISE Based Controller Development 525
Stability Analysis 527
Numerical Simulation 528
Conclusion 529
References 530
A Simplified Modular Petri Net for the Walking Assistant Robot 531
Introduction 531
Simplified Modular Petri Net (SMPN) 532
The Walkmate Robot System 533
Experiments 534
Conclusions 537
References 537
Omni-directional Vision Based Tracking and Guiding System for Walking Assistant Robot 539
Introduction 539
ODVS 540
Tracking 541
Marker and Marker Recognize Algorithm 541
Tracking Based on CA (Camshift Algorithm) 542
Occlusion and the Relative Position of Target 542
Guiding 542
Identification of Obstacles 543
APF Approach 543
Experiments 544
Tracking 544
Guiding 544
Conclusion 545
References 545
Dynamic Eye-in-Hand Visual Servoing with Unknown Target Positions 546
Introduction 546
Camera and Robot Model 547
Adaptive Image-Based Visual Servoing with Unknown Target Positions 548
Experiments 550
Conclusions 551
References 551
Optimum Motion Control for Stacking Robot 552
Introduction 552
Stacking Robot System Description 553
Application of Iterative Learning Control 554
Open-Loop Iterative Learning Control Theory 554
Application of Iterative Learning 554
Simulation Experiment and Result Analysis 556
Conclusion 559
References 559
Passive Target Tracking Using an Improved Particle Filter Algorithm Based on Genetic Algorithm 560
Introduction 560
Particle Filter and Drawbacks 561
Nonlinear Bayesian Filtering 561
Particle Filter 562
The Improved Particle Filter Algorithm 563
Simulation Experiments 564
Conclusion 566
References 567
Large-Scale Structure Assembly by Multiple Robots Which May Be Broken 568
Introduction 568
Space Solar Power Satellite 569
Overview of SSPS 569
SSPS Panel Module 569
SSPS Assembling Robot 570
Deadlock Problem 570
Leader-Follower Algorithm 571
Leader and Follower 571
Leader-Follower Exchange Method 572
Demand Map and Destination Decision 572
Simulation 573
Experimental Design 573
Evaluation Criteria 574
Simulation Result 1: Completion Rate 574
Simulation Result 2: Recovery Rate 575
Discussion 576
Conclusion 576
References 577
Real-Time Five DOF Redundant Robot Control Using a Decentralized Neural Scheme 578
Introduction 578
Discrete-Time Decentralized Systems 579
Neural Identifier 580
Decentralized Neural Identifier 580
EKF Training Algorithm 581
Controller Design 581
Five DOF Redundant Robot Application 583
Redundant Robot Description 584
Control Objective 584
Real-Time Results 585
Conclusions 587
References 587
Improving Transient Response of Adaptive Control Systems Using Multiple Neural Network Models 588
Introduction 588
Adaptive Control for Nonlinear System by Using Multiple RBF Neural Network Models 589
RBF Neural Network 589
Adaptive Control Using RBF Neural Network 590
Multiple Neural Network Model Adaptive Control 591
Step of Multiple Neural Network Model Adaptive Control 592
Simulation Analysis 593
Conclusions 594
References 595
An Information Theoretic Approach for Design MIMO Networked Control Systems 596
Introduction 596
Problem Formulation 596
Solution of Problem 598
Entropy of Tracking Error 598
MLP for Estimating the Pdf of Tracking Error 598
Controller Design 599
An Illustration Example 600
Conclusions 602
References 602
An Engineering Solution for Decoupling Control of Aircraft Motion Using Affine Neural Network 603
Introduction 603
Analysis of Longitudinal Flight Dynamics of a Light Airplane 603
Affine Neural Approximator and the Associated Training Algorithm 606
Input-Output Representation 606
Affine Neural Approximator 606
Affine Neural Approximation of Aerodynamic Coefficient Tables 607
Affine Neural Approximator for Drag and Pitching Moment Coefficients 608
Training of $N_{C_{D_{global}}}$ and $N_{C_{M_{local}}}$ 609
Simulation Results for Trajectory Tracking 611
Conclusion 612
References 613
Black-Box Input-Output Identification of a Class of Nonlinear Systems Using a Discrete-Time Recurrent Neurofuzzy Network 614
Introduction 614
Problem Statement 615
Proposed Network and Training Algorithm 617
Examples 618
Nonlinear Benchmark System 618
Experiment on a DC Motor 619
Conclusions 620
References 620
Passivity Analysis of Stochastic Neural Networks with Mixed Time-Varying Delays 622
Introduction 622
Problem Formulation and Preliminaries 623
Main Results 624
An Example 628
Conclusions 628
References 629
A Novel Recovering Initial Conditions Method from Spatiotemporal Complex Dynamical System 630
Introduction 630
Model and Method 631
The CD Research 632
The CD of Inverse Matrix of Coupling Coefficient Matrix 632
The CD of Inverse Mapping of Logistic Mapping 634
CD of the Inverse System of CML System 634
Simulation and Result Analysis 635
Conclusion 636
References 636
An Intelligent Control Scheme for Nonlinear Time-Varying Systems with Time Delay 638
Introduction 638
Smith Neural-Network Predictor (SNNP) 639
Parameter-Self-tuning Fuzzy Controller 641
Design of Fuzzy Controller 641
Self-tuning of the Scale Factor 642
Simulation 642
Simulation with Invariant System Parameters 642
Simulation with Time-Varying System Parameters 643
Conclusion 644
References 645
Master-Slave Chaos Synchronization of Uncertain Nonlinear Gyros Using Wavelet Neural Network 646
Introduction 646
Problem Statement 647
Description of WNN 648
AWNNC System Design 650
Simulation Results 651
Conclusions 653
References 653
Transportation Systems 22
WNN-Based Intelligent Transportation Control System 654
Introduction 654
Problem Formulation 655
Total Sliding-Mode Control (TSMC) System Design 655
Intelligent Transportation Control System (ITCS) Design 656
Illustrative Examples 658
Conclusions 659
References 660
Incident Detection in Urban Road 661
Introduction 661
The System Architecture 662
Experimental Results 666
Conclusions 668
References 668
An Efficient Web-Based Tracking System through Reduction of Redundant Connections 669
Introduction 669
System Architecture 671
Tracking Device 671
Tracking Server 672
Implementation 672
Conclusion 674
References 674
An Embedded All-Time Blind Spot Warning System 676
Introduction 676
System Overview 677
Day-Night Transmission Decision Module 678
Daytime Detection Module 679
Nighttime Detection Module 680
Distance Estimation Module and the Conclusions 680
Conclusions 681
References 681
Design of Autonomous Parallel Parking Using Fuzzy Logic Controller with Feed-Forward Compensation 683
Introduction 683
Parking Space Detection 685
Trajectory Following Control 685
Trajectory Planning 686
Trajectory Estimation 687
Fuzzy Logic Controller 687
Experimental Results 688
Conclusion 689
References 689
Telematics Services through Mobile Agents 691
Introduction 691
Telematics Services Delivering through OSGi-BasedMobile Agents 692
Mobile Agents in Contactless Smart Cards through a Java Bytecode Extractor 693
Information Filtering through a Risk-Enabled Reputation Model 694
Conclusion 696
References 697
Multi-agent System Model for Urban Traffic Simulation and Optimizing Based on Random Walk 699
Introduction 699
Modeling of System 700
The Proposed System 700
The Optimized Problem 703
Solution of the Optimized Problem 704
The Stable Solution of Random Walk 704
The Proposed System 705
Experiments and Results 705
Conclusion 707
References 707
Vehicle Detection Using Bayesian Enhanced CoBE Classification 708
Introduction 708
Bayesian Enhanced Cascades of Boosted Ensemble (BN+CoBE) Classification 709
Experiments 712
Conclusion 713
References 714
Vibration Analysis of a Submarine Model Based on an Improved ICA Approach 715
Introduction 715
Independent Component Analysis 716
Basic Theory of ICA 716
Separation Criterion Based on Negentropy 716
Improving the Separation Performance by Clustering Evaluation 717
Waveform Correlation Coefficient $/rho_{sy}$ 717
Simulation Experiment 717
Application to Quantitatively Calculate the Source Contributions of a Submarine Model 719
Conclusions 721
References 721
A Hierarchical Salient-Region Based Algorithm for Ship Detection in Remote Sensing Images 723
Introduction 723
Detection Algorithm 724
Preprocessing 725
Saliency Tile Extraction 726
Multiresource ROIs Extraction 728
Target Validation 729
Experimental Results and Discussion 730
Conclusions 731
References 731
Industrial Applications 23
Turning Tool Wear Monitoring Based on Fuzzy Cluster Analysis 733
Introduction 733
Fuzzy Clustering 734
Test-Bed Structures and Data Collection 734
Turning Data Acquisition Schematic 734
Test Conditions 735
Sensor and Installation 735
Tool Wear States Recognition 736
Signal Analysis 736
Feature Extraction 737
Fuzzy Clustering 738
Conclusions 739
References 739
Part-Machine Clustering: The Comparison between Adaptive Resonance Theory Neural Network and Ant Colony System 740
Introduction 740
Binary PMC Problems 741
Clustering Performance Measure 742
ART1 Neural Network for PMC Problems 743
How Does ART1 Work in Solving PMC Problems? 743
Numerical Example 744
ACS for PMC Problems 745
How Does ACS Work in Solving PMC Problems? 745
ACS State Transition Rule 745
ACS Local Pheromone Updating Rule 746
ACS Global Pheromone Updating Rule 746
Numerical Example 747
Conclusions and Future Work 748
References 748
Fault Diagnosis of Bearings Based on Time-Delayed Correlation and Demodulation as Well as B-Spline Fuzzy Neural Networks 749
Introduction 749
The Fault Diagnosis System for Bearings and Data Collection 750
Time-Delayed Correlation and Demodulation Technique 750
The Basic Theory 750
The Application of Time-Delayed Correlation and Demodulation Technique 751
The B-Spline Neurofuzzy Networks and Its Application 752
Bearing Fault Recognition by Using B-Spline Neurofuzzy Networks 752
Conclusion 754
References 755
Fast and Noninvasive Determination of Viscosity of Lubricating Oil Based on Visible and Near Infrared Spectroscopy 756
Introduction 756
Materials and Methods 757
Instrument Settlement 757
Sample Preparation and Spectral Collection 757
Spectral Preprocessing 758
Calibration Methods 758
Results and Discussion 759
Extraction of Input Eigenvectors 759
MLR and PLS Models 760
BPNN Model 760
Model Comparison 761
Conclusion 761
References 762
Chattering-Free Adaptive Wavelet Neural Network Control for a BLDC Motor via Dynamic Sliding-Mode Approach 763
Introduction 763
Problem Statement 764
CAWNNC System Design 765
Description of WNN 765
Design of CAWNNC 766
Experimental Results 768
Conclusions 770
References 770
A New BPSO Algorithm and Applications in Interruptible Load Management 772
Introduction 772
The Multi-objective Programming Model of ILS 773
Constraints 773
Some Definitions 774
Multi-objective Natural Learning Binary Particle Swarm Optimization (MNLBPSO ) Algorithm 775
Binary Particle Swarm Optimization (BPSO) Algorithm 775
Natural Learning Binary Particle Swarm Optimization Algorithm 776
Experiment Results 777
Conclusions 779
References 779
Force Identification by Using Support Vector Machine and Differential Evolution Optimization 781
Introduction 781
Method 782
Support Vector Machine 782
Differential Evolution (DE) [10] 783
SVM-DE Hybrid Model 784
Numerical Cases and Experiment Study 785
Numerical Simulations 785
Experiment Study 787
Conclusions 788
References 789
Soft Sensor Modeling of Ball Mill Load via Principal Component Analysis and Support Vector Machines 790
Introduction 790
Scheme of the Grinding Process and Mill Load 791
Soft Sensor Modeling of Mill Load 792
Soft Sensor Strategy 792
Principal Component Analysis 793
Support Vector Machines 794
Steps of Soft Sensor Algorithms 795
Application Study 795
Conclusions 797
References 797
An Approach Based on Hilbert-Huang Transform and Support Vector Machine for Intelligent Fault Diagnosis 798
Introduction 798
Features Extraction Based on HHT 799
The Hilbert Spectrum and Marginal Spectrum 799
Features Extraction 800
Application of LS-SVM in Fault Diagnosis of Tool Cutting Based on HHT 802
Conclusions 803
References 804
Real-World Applications 24
Study on Factors of Floating Women’s Income in Jiangsu Province Based on Bayesian Networks 805
Introduction 805
The Data Set 806
Bayesian Network 806
Introduction 806
Structure Learning Algorithms 807
Results 808
Network Scores 810
Conclusion 812
References 813
Variation Trend Analysis of Groundwater Depth inArea of Well Irrigation in Sanjiang Plain Based on Wavelet Neural Network 814
Introduction 814
Wavelet Analysis 815
Wavelet Transform 815
Fast Wavelet Transform Algorithm 815
Wavelet Neural Network 816
Basic Principle 816
Model Structure 816
Main Steps of Modeling 816
Case Study 817
Wavelet Decomposition and Reconstruction of Measured Series 817
Determination of Input and Output Paired Samples 817
Determination of WANN Model Structure 818
WANN Model Fitting 818
Precision Inspection of WANN Model 819
Prediction of Groundwater Depth 819
Discussions 820
Conclusions 820
References 821
A Petri-Net Modeling Method of Agent’s Belief-Desire-Intention and Its Application in Logistics 822
Introduction 822
Petri-Net Modeling Method of Agent’s Belief-Desire-Intention 823
Modeling Inventory Systems with BDIPNs and Batch Deterministic and Stochastic Petri Nets 825
Application of BDI-BDSPN in a Complex Logistics 827
Application of BDI-BDSPN in an Agile Logistics 828
Conclusion 829
References 829
Supply Chain Flexibility Assessment by Multivariate Regression and Neural Networks 830
Introduction 830
Selected Supply Chain Flexibility Dimensions 831
Multiple Regression Analysis 832
Neural Network Simulation 834
Discussion and Conclusion 836
References 836
An Intelligent Mobile Location-Aware Book Recommendation System with Map-Based Guidance That Enhances Problem-Based Learning in Libraries 838
Introduction 838
System Design 839
System Architecture 839
The Employed Average-KNN Classifier for Identifying Learner Location 840
Recommending Books Associated with Learner Search Query and Location 841
The Implemented Intelligent Mobile Location-Aware Book Recommendation System with Map-Based Guidance for Supporting PBL 841
Experiment Design 842
Experimental Results 842
WLAN Positioning Experiment 842
Learning Performance Assessment 843
Conclusion 844
References 845
Applying Least Squares Support Vector Regression with Genetic Algorithms for Radio-Wave Path-Loss Prediction in Suburban Environment 846
Introduction 847
Egli Model 847
Walfisch-Bertoni Model 847
Least Square Support Vector Regression with Genetic Algorithm 848
Numerical Examples and Experimental Results 850
Conclusions 852
References 853
A Probabilistic Neural Network Approach to Modeling the Impact of Tobacco Control Policies by Gender 854
Introduction 854
The International Tobacco Control Survey Data on Smokers’ Quitting Motivations 855
Identifying Tobacco Control Policy Drivers 856
Building Probabilistic Neural Network Models Using Policy Driver Attributes 857
Probabilistic Neural Network 857
Developing Probabilistic Neural Network Models for Groups of Female and Male Smokers 858
Evaluating the Impact of Tobacco Control Policies on Female and Male Smokers 858
Discussion 859
Conclusion 860
References 860
The BPNN-Fuzzy Logic Pre-control of an Underground Project in City Center of Shanghai 862
Introduction 862
Analysis of the Foundation Pit Displacement Data during Prophase Excavation 864
Analysis and Prediction on the Horizontal Displacements of Project Foundation Pit by BP Neural Networks 865
The Fuzzy Logic Prediction of the Foundation Pit Excavation Control Measures of Building No.4 866
Conclusions 868
References 869
Optimal Parameter Inversion of Marine Water Quality Model Using a BPNN Data-Driven Model –– A Case Study on DIN 870
Introduction 870
Numerical Model 871
BPNN Data-Driven Model 871
Water Quality Model 872
Optimal Parameters Inversion Model 872
Case Studies 872
Choices of Control Variable 874
Optimal Estimation 874
Verification of Optimal Solution 876
Conclusion 876
References 876
Determination of Sugar Content of Instant Milk-Tea Using Effective Wavelengths and Least Squares-Support Vector Machine 878
Introduction 878
Materials and Methods 879
Instant Milk-Tea Samples 879
Spectral Acquisition and Pretreatment 879
Partial Least Squares Analysis 880
Least Squares-Support Vector Machine 881
Results and Discussion 881
Spectral Features and Statistics of Sugar Content 881
PLS Models and Analysis of LVs and EWs 882
LS-SVM Models 882
Conclusion 884
References 885
Sports Video Summarization Based on Salient Motion Entropy and Information Analysis 886
Introduction 886
Related Work 887
Proposed Method 887
Saliency Map Extraction 888
Salient Motion Entropy 888
Mutual Information Based on Salient Motions 889
Experimental Results 890
Conclusion 891
References 892
A Neural Network Based Algorithm for the Retrieval of Precipitable Water Vapor from MODIS Data 894
Introduction 894
Algorithm 895
Methodology 895
Neural Network 896
Training and Testing of the MFNN 896
Validation and Discussion 897
Validation of the MFNN Algorithm Using MODIS Data 897
Validation of the MFNN Algorithm Using Radiosonde Data 897
Discussion 898
Conclusion 900
References 900
A Neural Network Based Approach to Wind Energy Yield Forecasting 902
Introduction 902
Neural Network Design 903
Neural Network Output 904
Wind Spectra Computation 905
Turbine Output Calculations 906
Turbine Output Calculations 907
Errors Rates in Estimations 907
Conclusion 908
References 909
Research on New Intelligent Business-Oriented Decision-Making Model Based on MA and GA 910
Introduction 910
The Solution Method Based on Cognition 911
Model and Algorithm 913
Conclusions 916
References 916
Author Index 918

"A Novel Prediction Mechanism with Modified Data Mining Technique for Call Admission Control in Wireless Cellular Network (S. 1-2)

Abstract. It is an important issue to allocate appropriate resources to mobile calls for wireless cellular networks owe to scarce wireless spectrums. The call admission control (CAC) will maintain better performance metrics of mobile call such as call dropping probability (CDP) and call blocking probability (CBP) if the future utilization of wireless spectrums can be predicted and provided to the decision of CAC. Therefore, a prediction mechanism which can predict most information such as system utilization is proposed in this paper.

The techniques of data mining and pattern matching which adopts gradient to fuzz time series data for representations of chain code are applied to mining a possible repetitive pattern. Our proposed prediction mechanism can provide prediction information in advance whether the repetitive time series pattern of information exists or not. Furthermore, an update of confident level will be conducted continuously for performing each prediction in the proposed scheme.

Our proposed mechanism is developed and tested with four cases which can be regarded as using scenarios of wireless cellular network. The experimental results show that the proposed scheme can capture repetitive time series patterns and perform following predictions with these repetitive time series patterns. Besides, the required storage is less than traditional schemes and lower computation power is required for the proposed scheme. Keywords: Call Admission Control (CAC), data mining, time series, pattern matching, prediction.

1 Introductions

Although there has been a rapid development in wireless cellular communications, the QoS guarantee remains one of the most challenging issues [1]. One of the key elements in providing QoS guarantees is an effective CAC policy, which not only has to ensure that the network meets the QoS of the newly arriving calls if accepted, but also guarantees that the QoS of the existing calls does not deteriorate. The variable user mobility has made that it becomes more complex to predict the appropriate cell for handoff.

The past research [2] showed the impact of mobility on cellular network and provided a modeling method for configuring cellular networks to study the dynamics of mobility. The improvement of radio bandwidth is always thought as a dynamic channel (code) allocation problem in literatures [3]. Although there were some schemes for bandwidth reservation proposed to reduce the CDP, such as the study [4], seldom literature has developed to satisfy QoS and lower CBP issues at the same time. Furthermore, the past researches mentioned at the above focused on individual mobility prediction, and they may cause mass load focusing on the MSC.

According to the study [5], final information is a prediction of users’ number in a given cell, and it leads to use a global approach that only observes variations of system utilization and users’ flows. There are many advantages for [5]: it does not require any control message and additional load for the MSC; cells generate their own statistics independently from others; it is sensitive to geographical constrains and to users’ common habits. According to [12], the final information is a quantity prediction of users in a given cell, and it leads to use a global approach that only observes variations of system utilization and users’ flows.

There are many advantages of the scheme presented in [12]: it does not require any control message and additional load for the MSC; cells generate their own statistics independently from others; it is sensitive to geographical constraints and to users’ common habits. Besides, the concept of aggregated history has been also applied to acquire user mobility profile in [13] so that a user mobility profile framework is developed for estimating service patterns and tracking mobile users, including descriptions of location, mobility, and service requirements. Therefore, in order to provide suitable statistical prediction information which may be system utilization, CBP, or other system resources to the CAC, a prediction mechanism which can predict most information is proposed in this paper."