最大相关熵准则自适应滤波器的分数阶长算法

刘春辉; 齐越; 丁文锐; 张文秋

doi:10.13700/j.bh.1001-5965.2015.0137

最大相关熵准则自适应滤波器的分数阶长算法

doi: 10.13700/j.bh.1001-5965.2015.0137

刘春辉^1,2,
齐越²,
丁文锐^1, ,,
张文秋³

1.
北京航空航天大学无人驾驶飞行器设计研究所, 北京 100083
2. 北京航空航天大学计算机学院, 北京 10008
3. 北京航空航天大学电子信息工程学院, 北京 100083

基金项目: 国家自然科学基金(61450008,61301004)

详细信息

作者简介:
刘春辉男,博士研究生,工程师。主要研究方向:数据链、自适应信号处理和图像融合。Tel.:010-82339906 E-mail:liuchunhui2134@126.com;丁文锐女,博士,研究员。主要研究方向:数据链、计算机视觉和人工智能。Tel.:010-82339906 E-mail:ding@buaa.edu.cn

通讯作者:
丁文锐,Tel.:010-82339906 E-mail:ding@buaa.edu.cn

中图分类号: TN911.72
计量
- 文章访问数: 1185
- HTML全文浏览量: 47
- PDF下载量: 569
- 被引次数: 0
出版历程
- 收稿日期: 2015-03-12
- 网络出版日期: 2016-02-20

Fractional tap-length algorithm for adaptive filters under maximum correntropy criterion

1.
Research Institute of Unmanned Aerial Vehicle, Beijing University of Aeronautics and Astronautics, Beijing 100083, China
2. School of Computer Science and Engineering, Beijing University of Aeronautics and Astronautics, Beijing 10008
3. School of Electronic and Information Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100083, China

摘要

摘要: 对自适应滤波器最佳阶长的准确估计可以有效平衡自适应算法的稳态性能与计算复杂度,而基于最小均方差(MMSE)准则的变阶长最小均方(LMS)算法在非高斯噪声环境下的收敛性能变差。针对这一问题,提出一种最大相关熵准则(MCC)自适应滤波器的分数阶长(FT)算法——FT-MCC算法。该算法从MCC自适应滤波器最佳阶长的定义出发,利用不同阶长滤波器产生的相关熵之差实现阶长更新。理论分析和实验表明:相比现有变阶长最小均方算法,FT-MCC算法在非高斯噪声环境中具有较强的鲁棒性;通过恰当的参数选择,算法可较好地实现对最佳阶长的跟踪和估计。
- 自适应滤波 /
- 最大相关熵准则 /
- 变阶长算法 /
- 分数阶长算法 /
- 系统辨识 /
- 鲁棒性
Abstract: Accurate estimation of the optimum tap-length for the adaptive filter provides a good balance between the steady-state performance and the complexity of the adaptive algorithm. The convergence performance of the variable tap-length least mean square (LMS) adaptive filters under the minimum mean square error (MMSE) criterion deteriorates in the non-Gaussian noise environment. A fractional tap-length (FT) algorithm for the maximum correntropy criterion (MCC) adaptive filters, named FT-MCC algorithm, is proposed to solve the above problem. The proposed algorithm is based on the concept of the optimum tap-length for the MCC adaptive filters. The difference of the correntropy between adaptive filters with different tap-lengths is used to achieve the tap-length update. Both theoretical analysis and simulation result show that the proposed algorithm has strong robustness in non-Gaussian noise environment compared with other variable tap-length LMS algorithm and the optimum tap-length can be well estimated with proper parameter selection.
- adaptive filtering /
- maximum correntropy criterion /
- variable tap-length algorithm /
- fractional tap-length algorithm /
- system identification /
- robustness

HTML全文

参考文献(19)

[1]	TAN Y,HE Z Q,TIAN B Y.A novel generalization of modified LMS algorithm to fractional order[J].IEEE Signal Processing Letters,2015,22(9):1244-1248.
[2]	SCHOBER R,GERSTACKER W H,LAMPE A.Non-coherent MMSE interference suppression for DS-CDMA[J].IEEE Transactions on Communications,2002,50(4):577-587.
[3]	SLOCK D T M.On the convergence behavior of the LMS and the normalized LMS algorithms[J].IEEE Transactions on Signal Processing,1993,41(9):2811-2825.
[4]	RIERA-PALOU F,NORAS J M,CRUICKSHANK D G M.Linear equalizers with dynamic and automatic length selection[J].Electronics Letters,2001,37(25):1553-1554.
[5]	GU Y T,TANG K,CUI H J.LMS algorithm with gradient descent filter length[J].IEEE Signal Processing Letters,2004,11(3):305-307.
[6]	GONG Y,COWAN C F N.An LMS style variable tap-length algorithm for structure adaptation[J].IEEE Transactions on Signal Processing,2005,53(7):2400-2407.
[7]	LI L L,CHAMBERS J A.A novel adaptive leakage factor scheme for enhancement of a variable tap-length learning algorithm[C]//IEEE International Conference on Acoustics,Speech and Signal Processing.Piscataway,NJ:IEEE Press,2008:3837-3840.
[8]	XU D J,YIN B,WANG W,et al.Variable tap-length LMS algorithm based on adaptive parameters for TDL structure adaption[J].IEEE Signal Processing Letters,2014,21(7):809-813.
[9]	LIU Z Y.Variable tap-length linear equaliser with variable tap-length adaptation step-size[J].Electronics Letters,2014,50(8):587-589.
[10]	ERDOGMUS D,PRINCIPE J C.Generalized information potential criterion for adaptive system training[J].IEEE Transactions on Neural Networks,2002,13(5):1035-1044.
[11]	SINGH A,PRINCIPE J C.Using correntropy as a cost function in linear adaptive filters[C]//International Joint Conference on Neural Networks.Piscataway,NJ:IEEE Press,2009:2950-2955.
[12]	LIU W,POKHAREL P P,PRINCIPE J C.Correntropy:Properties and applications in non-Gaussian signal processing[J].IEEE Transactions on Signal Processing,2007,55(11):5286-5298.
[13]	CHEN B,LEI X,LIANG J,et al.Steady-state mean-square error analysis for adaptive filtering under the maximum correntropy criterion[J].IEEE Signal Processing Letters,2014,21(7):880-884.
[14]	HE R,ZHENG W S,HU B G.Maximum correntropy criterion for robust face recognition[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,2011,33(8):1561-1576.
[15]	ZHAO S,CHEN B,PRINCIPE J C.Kernel adaptive filtering with maximum correntropy criterion[C]//Proceedings of International Joint Conference on Neural Networks.Piscataway,NJ:IEEE Press,2011:2012-2017.
[16]	WEI X H,CRUICKSHANK D G M,MULGREW B,et al.A unified approach to dynamic length algorithms for adaptive linear equalizers[J].IEEE Transactions on Signal Processing,2007,55(3):908-920.
[17]	ZHANG Y,LI N,CHAMBER J A,et al.Steady-state performance analysis of a variable tap-length LMS algorithm[J].IEEE Transactions on Signal Processing,2008,56(2):839-845.
[18]	HAYKIN S.Adaptive filter theory[M].4th ed.Englewood Cliffs,NJ:Prentice Hall Press,1996:276-277.
[19]	刘明骞,李兵兵,曹超凤.非高斯噪声下数字调制信号识别方法[J].电子与信息学报,2013,35(1):85-91. LIU M Q,LI B B,CAO C F.Recognition method of digital modulation signals in non Gaussian noise[J].Journal of Electronics & Information Technology,2013,35(1):85-91(in Chinese).