Performance improvement of the NLMS algorithm and its applications 学習同定法の性能改善とその応用
この論文にアクセスする
この論文をさがす
著者
書誌事項
- タイトル
-
Performance improvement of the NLMS algorithm and its applications
- タイトル別名
-
学習同定法の性能改善とその応用
- 著者名
-
Jirasak Tanpreeyachaya
- 著者別名
-
ジラサック タンパリヤチャヤー
- 学位授与大学
-
名古屋工業大学
- 取得学位
-
博士 (工学)
- 学位授与番号
-
甲第167号
- 学位授与年月日
-
1996-03-22
注記・抄録
博士論文
目次
- Contents / p1 (0003.jp2)
- 1 Introduction / p1 (0007.jp2)
- 1.1 The Linear Filtering Problem / p1 (0007.jp2)
- 1.2 Adaptive Filters / p2 (0008.jp2)
- 1.3 Applications / p5 (0009.jp2)
- 1.4 Outline of the Thesis / p7 (0010.jp2)
- 2 Adaptive Systems and Theory of Adaptation / p11 (0012.jp2)
- 2.1 General Description / p11 (0012.jp2)
- 2.2 Input Signal and Weight Vectors / p11 (0012.jp2)
- 2.3 Desired Response and Error / p13 (0013.jp2)
- 2.4 The Performance Function / p15 (0014.jp2)
- 2.5 Gradient and Minimum Mean-square Error / p17 (0015.jp2)
- 2.6 Alternative Expression of the Gradient / p18 (0016.jp2)
- 2.7 Properties of the Quadratic Performance Surface / p20 (0017.jp2)
- 2.8 Normal Form of the Input Correlation Matrix / p20 (0017.jp2)
- 2.9 Eigenvalues and Eigenvectors of the Input Correlation Matrix / p21 (0017.jp2)
- 2.10 Geometrical Significance of Eigenvectors and Eigenvalues / p23 (0018.jp2)
- 2.11 Searching the Performance Surface / p26 (0020.jp2)
- 2.12 Basic Ideas of Gradient Search Methods / p27 (0020.jp2)
- 2.13 Simple Gradient Search and Its Solution / p28 (0021.jp2)
- 2.14 Stability and Rate of Convergence / p29 (0021.jp2)
- 2.15 The Learning Curve / p30 (0022.jp2)
- 2.16 Gradient Search by Newton's Method / p31 (0022.jp2)
- 2.17 Newton's Method in Multidimensional Space / p34 (0024.jp2)
- 2.18 Gradient Search by the Method of Steepest Descent / p35 (0024.jp2)
- 3 The LMS and NLMS Algorithm / p41 (0027.jp2)
- 3.1 Derivation of the LMS algorithm / p41 (0027.jp2)
- 3.2 Derivation of the NLMS algorithm / p47 (0030.jp2)
- 4 The Conditioned NLMS method / p51 (0032.jp2)
- 4.1 Introduction / p51 (0032.jp2)
- 4.2 The Conditioned NLMS Algorithm / p52 (0033.jp2)
- 4.3 Steady State Residual Error Analysis / p54 (0034.jp2)
- 4.4 Transient State Length Analysis / p59 (0036.jp2)
- 4.5 Optimal Selection of the Threshold p / p63 (0038.jp2)
- 4.6 Comparison with the ε-NLMS Method / p68 (0041.jp2)
- 4.7 Conclusion / p74 (0044.jp2)
- 5 The Partial-NLMS method / p75 (0044.jp2)
- 5.1 Introduction / p75 (0044.jp2)
- 5.2 Partial Normalized LMS Algorithm / p76 (0045.jp2)
- 5.3 Convergence Behavior Analysis / p77 (0045.jp2)
- 5.4 Performance Comparison between the P-NLMS and NLMS Algorithm / p82 (0048.jp2)
- 5.5 Application of P-NLMS to an IIR ADF / p83 (0048.jp2)
- 5.6 Conclusion / p87 (0050.jp2)
- 6 Performance Improvement of Variable Stepsize NLMS / p89 (0051.jp2)
- 6.1 Introduction / p89 (0051.jp2)
- 6.2 Optimal Controlled Stepsize NLMS / p92 (0053.jp2)
- 6.3 The Noise Variance Estimation Method and VS-NLMS Algorithm / p93 (0053.jp2)
- 6.4 Simulation Results / p98 (0056.jp2)
- 6.5 Conclusions / p102 (0058.jp2)
- 7 Summary / p113 (0063.jp2)
- A Derivation of the Correlation Eq.(4.5) / p115 (0064.jp2)
- B Derivation of Eq.(4.10) / p117 (0065.jp2)
- C Derivation of Eq.(4.11) and (4.12) / p119 (0066.jp2)
- D Proof of the Condition for[数式]<1 and[数式]≈0.5N / p121 (0067.jp2)
- D.1 Sufficient Condition for[数式]<1 / p121 (0067.jp2)
- D.2 Consideration for[数式]≈O.5N / p122 (0068.jp2)
- E Derivation of Eq.(5.8),(5.9) and (5.10) / p123 (0068.jp2)
- Acknowledgments / p125 (0069.jp2)
- Reference / p127 (0070.jp2)
- Authors'published papers / p130 (0072.jp2)
- Biography / p131 (0072.jp2)