New normalized LMS adaptive filter with a variable regularization factor

doi:10.21629/JSEE.2019.02.05

Journal of Systems Engineering and Electronics ›› 2019, Vol. 30 ›› Issue (2): 259-269.doi: 10.21629/JSEE.2019.02.05

收稿日期:2018-06-07 出版日期:2019-04-01 发布日期:2019-04-28

New normalized LMS adaptive filter with a variable regularization factor

Zhoufan LI(), Dan LI*(), Xinlong XU(), Jianqiu ZHANG()

School of Information Science and Technology, Fudan University, Shanghai 200433, China

Received:2018-06-07 Online:2019-04-01 Published:2019-04-28
Contact: Dan LI E-mail:15210720031@fudan.edu.cn;lidan@fudan.edu.cn;11210720032@fudan.edu.cn;jqzhang@ieee.org
About author:LI Zhoufan was born in 1992. He received his B.Sc. degree in electrical engineering from Fudan University, Shanghai, China, in 2015. He is currently working towards his M.S. degree in Department of Electronic Engineering, Fudan University, Shanghai, China. His research interests include digital signal processing and its application in wireless communication. E-mail:15210720031@fudan.edu.cn|LI Dan was born in 1982. He received his B.Sc., M.S., and Ph.D. degrees in electrical engineering from Fudan University, Shanghai, China, in 2003, 2006, and 2013, respectively. Since 2006, he has been an lecturer at Department of Electronic Engineering, Fudan University, Shanghai, China. His research interests include digital signal processing and its application to nondestructive testing. E-mail:lidan@fudan.edu.cn|XU Xinlong was born in 1989. He received his B.Sc. and M.S. degrees in electrical engineering from Fudan University, Shanghai, China, in 2011 and 2014, respectively. His research interests include signal processing and adaptive filtering. E-mail:11210720032@fudan.edu.cn|ZHANG Jianqiu was born in 1962. He received his B.Sc. degree from East of China Institute of Engineering, Nanjing, and his M.S. and Ph.D. degrees from Harbin Institute of Technology (HIT), Harbin, China, in 1992 and 1996, respectively. He is currently a professor with the Department of Electronic Engineering, Fudan University, Shanghai, China. From 1999 to 2002, he was a senior research fellow at the School of Engineering, University of Greenwich, Chatham Maritime, U.K. In 1998, he was a visiting research scientist at the Institute of Intelligent Power Electronics, Helsinki University of Technology, Espoo, Finland. He was an associate professor from 1995 to 1997 and a lecturer from 1989 to 1994 in the Department of Electrical Engineering, HIT. During 1982 to 1987, he was an assistant electronic engineer at the 544th Factory, Hunan, China. His main research interests are signal processing and its application. Email:jqzhang@ieee.org
Supported by:
the National Natural Science Foundation of China(61571131);the National Natural Science Foundation of China(11604055);This work was supported by the National Natural Science Foundation of China (61571131; 11604055)

摘要/Abstract

Abstract:

A new normalized least mean square (NLMS) adaptive filter is first derived from a cost function, which incorporates the conventional one of the NLMS with a minimum-disturbance (MD) constraint. A variable regularization factor (RF) is then employed to control the contribution made by the MD constraint in the cost function. Analysis results show that the RF can be taken as a combination of the step size and regularization parameter in the conventional NLMS. This implies that these parameters can be jointly controlled by simply tuning the RF as the proposed algorithm does. It also demonstrates that the RF can accelerate the convergence rate of the proposed algorithm and its optimal value can be obtained by minimizing the squared noise-free posteriori error. A method for automatically determining the value of the RF is also presented, which is free of any prior knowledge of the noise. While simulation results verify the analytical ones, it is also illustrated that the performance of the proposed algorithm is superior to the state-of-art ones in both the steady-state misalignment and the convergence rate. A novel algorithm is proposed to solve some problems. Simulation results show the effectiveness of the proposed algorithm.

Key words: adaptive filtering, normalized least mean square (NLMS), minimum-disturbance (MD) constraint, variable regularization, variable step-size NLMS

. [J]. Journal of Systems Engineering and Electronics, 2019, 30(2): 259-269.

Zhoufan LI, Dan LI, Xinlong XU, Jianqiu ZHANG. New normalized LMS adaptive filter with a variable regularization factor[J]. Journal of Systems Engineering and Electronics, 2019, 30(2): 259-269.

图/表 3

参考文献 33

1	WIDROW B, STEARNS S D. Adaptive signal processing. Englewood Cliffs: Prentice-Hall, 1985.
2	HAYKIN S. Adaptive filter theory. 4th ed. Upper Saddle Rier, New Jersey: Prentice-Hall, 2002.
3	BENESTY J, HUANG Y. Adaptive signal processingapplications to real-world problems. Berlin: Springer-Verlag, 2003.
4	HUANG Y, BENESTY J, CHEN J. Acoustic MIMO signal processing. Boston, MA: Springer, 2006.
5	MA T, CHEN J, CHEN W, et al. Unsupervised robust adaptive filtering against impulsive noise. Journal of Systems Engineering and Electronics, 2012, 23 (1): 32- 39. doi: 10.1109/JSEE.2012.00005
6	GUO L, JUNG L. Necessary and sufficient conditions for stability of LMS. IEEE Trans. on Automatic Control, 1995, 42 (6): 761- 770.
7	HASSIBI B, SAYED A H, KAILATH T. H^∞ optimality of the LMS algorithm. IEEE Trans. on Signal Processing, 1996, 44 (2): 267- 280. doi: 10.1109/78.485923
8	EVANS J B, XUE P, LIU B. Analysis and implementation of variable step size adaptive algorithms. IEEE Trans. on Signal Processing, 1993, 41 (8): 2517- 2535. doi: 10.1109/78.229885
9	ABOULNASR T, MAYYAS K. A robust variable step-size LMS-type algorithm:analysis and simulations. IEEE Trans. on Signal Processing, 1997, 45 (3): 631- 639. doi: 10.1109/78.558478
10	PAZAITIS D I, CONSTANTINIDES A G. A novel kurtosis driven variable step-size adaptive algorithm. IEEE Trans. on Signal Processing, 2002, 47 (3): 864- 872.
11	ANTWEILER C, GRUNWALD J, QUACK H. Approximation of optimal step size control for acoustic echo cancellation. Proc. of the IEEE International Conference on Acoustics, 1997, 295- 298.
12	MORGAN D R, KRATZER S G. On a class of computationally efficient, rapidly converging, generalized NLMS algorithms. IEEE Signal Processing Letters, 1996, 3 (8): 245- 247. doi: 10.1109/97.511808
13	GOLLAMUDI S, NAGARAJ S, KAPOOR S, et al. Setmembership filtering and a set-membership normalized LMS algorithm with an adaptive step size. IEEE Signal Processing Letters, 1998, 5 (5): 111- 114. doi: 10.1109/97.668945
14	SHIN H C, SAYED A H, SONG W J. Variable step-size NLMS and affine projection algorithms. IEEE Signal Processing Letters, 2008, 11 (2): 132- 135.
15	BENESTY J, REY H, VEGA L R, et al. A nonparametric VSS NLMS algorithm. IEEE Signal Processing Letters, 2006, 13 (10): 581- 584. doi: 10.1109/LSP.2006.876323
16	HUANG H C, LEE J. A new variable step-size NLMS algorithm and its performance analysis. IEEE Trans. on Signal Processing, 2012, 60 (4): 2055- 2060. doi: 10.1109/TSP.2011.2181505
17	ARABLOUEI R, DOGANCAY K. Linearly-constrained recursive total least-squares algorithm. IEEE Signal Processing Letters, 2012, 19 (12): 821- 824. doi: 10.1109/LSP.2012.2221705
18	BENESTY J, GAENSLER T, MORGAN D R, et al. Advances in network and acoustic echo cancellation. Berlin: SpringerVerlag, 2001.
19	FAZA A, GRANT S L, BENESTY J. Adaptive regularization in frequency-domain NLMS filters. Proc. of the Signal Processing Conference, 2012, 2625- 2628.
20	BENESTY J, PALEOLOGU C, CIOCHINA S. On regularization in adaptive filtering. IEEE Trans. on Audio, Speech and Language Processing, 2011, 19 (6): 1734- 1742. doi: 10.1109/TASL.2010.2097251
21	NI J, LI F. Variable regularisation parameter sign subband adaptive filter. Electronics Letters, 2010, 46 (24): 1605- 1607. doi: 10.1049/el.2010.2406
22	MADER A, PUDER H, SCHMIDT G U. Step-size control for acoustic echo cancellation filters-an overview. Signal Processing, 2000, 80 (9): 1697- 1719. doi: 10.1016/S0165-1684(00)00082-7
23	SULYMAN A I, ZERGUINE A. Convergence and steady-state analysis of a variable step-size NLMS algorithm. Signal Processing, 2003, 83 (6): 1255- 1273. doi: 10.1016/S0165-1684(03)00044-6
24	PARK P, CHANG M, KONG N. Scheduled-stepsize NLMS algorithm. IEEE Signal Processing Letters, 2009, 16 (12): 1055- 1058. doi: 10.1109/LSP.2009.2026197
25	SONG I, PARK P G. A normalized least-mean-square algorithm based on variable-step-size recursion with innovative input data. IEEE Signal Processing Letters, 2012, 19 (12): 817- 820. doi: 10.1109/LSP.2012.2221699
26	YUAN Z, SONGTAO X. New LMS adaptive filtering algorithm with variable step size. Proc. of the International Conference on Vision, Image and Signal Processing, 2017, 1- 4.
27	LI M, XI X. A new variable step-size NLMS adaptive filtering algorithm. Proc. of the Information Technology and Applications, 2013, 236- 239.
28	HUAN Q Y, QIU X H, LIU X F. Variable step LMS algorithm using norm of the hyperbolic tangent function. Journal of Signal Processing, 2014, 30 (1): 93- 99.
29	WANG Y, SUN X, LIU L. A variable step size LMS adaptive filtering algorithm based on L2 norm. Proc. of the Signal Processing, Communications and Computing, 2016, 1- 6.
30	CIOCHINĂ S, PALEOLOGU C, BENESTY J. An optimized NLMS algorithm for system identification. Signal Processing, 2016, 118, 115- 121. doi: 10.1016/j.sigpro.2015.06.016
31	BHOTTO M Z A, ANTONIOU A. New improved recursive least-squares adaptive-filtering algorithms. IEEE Trans. on Circuits and Systems Ⅰ:Regular Papers, 2013, 60 (6): 1548- 1558. doi: 10.1109/TCSI.2012.2220452
32	CASTOLDI F T, DE CAMPOS M L R. Application of a minimum-disturbance description to constrained adaptive filters. IEEE Signal Processing Letters, 2013, 20 (12): 1215- 1218. doi: 10.1109/LSP.2013.2284384
33	GUL M M U, MA X, LEE S. Timing and frequency synchronization for OFDM downlink transmissions using zadoff-chu sequences. IEEE Trans. on Wireless Communications, 2015, 14 (3): 1716- 1729. doi: 10.1109/TWC.2014.2372757

Algorithm	Variable step	Parameter
SM-NLMS [13]	$\mu (n)=\left\{\!\!\begin{array}{*{20}l} {1-\dfrac{\gamma }{\|e(n)\|}}, ~~~ {\|e(n)\|>\gamma } \\ 0, ~~~ {\rm else}\\ \end{array}\!\!\right.$	$\gamma $ = 0.04 ${\ldots}$
NPVSS-NLMS [15]	$\mu(n)=\left\{\!\!\begin{array}{*{20}l} {1-\dfrac{\sigma _v }{\widehat {\sigma }_e (n)+\varepsilon }}, ~~~ {\widehat {\sigma }_e (n)\ge \widehat {\sigma }_e (n)} \\ 0, ~~~ {\rm else} \\ \end{array}\!\!\right.$ $\widehat {\sigma }_e (n)=\lambda \widehat {\sigma }_e (n-1)+(1-\lambda)e^2(n)$	$\lambda = 1-1/(LK)$, $\delta = 20\sigma _{x}^{2}$, $\varepsilon = 10^{-3}$, $K = 2$ for a white input and $K = 6$ for an AR(1) process. $\sigma _{v}^{2}$ and $\sigma _{x}^{2}$ are known as the priori.
NVSS-NLMS [16]	$\mu (n)=\left\{\!\!\begin{array}{*{20}l} \left[\alpha \mu (n-1)+(1-\alpha)\dfrac{\widehat {\sigma }_e^2 (n)}{\beta \widehat {\sigma }_v^2 (n)} \right] \bigg\|_{\mu _{\min } }^{\mu _{\max } }, ~~~ {\zeta (n) < \zeta _{th} } \\ 0, ~~~ {\rm else} \\\end{array}\!\!\right.$ $\widehat {\sigma }_e^2 (n)=\alpha \widehat {\sigma }_e^2 (n-1)+(1-\alpha)e^2(n)$	$\alpha = 0.998$, $\beta = 30$, $\zeta _{\rm th} = 0.35$, $\mu_{\min} = 10^{-5}$, and $\mu_{\max} = 1$.
JO-NLMS [30]	$q(n)=\dfrac{p(n)}{L\sigma _v^2 +(L+2)p(n)\widehat {\sigma }_x^2 }$	$\sigma _{v}^{2}$ is known as a priori.
NVSS-LMS [26]	$\mu (n)=\beta \{1/[1+\exp (-\alpha \|e(n)e(n-1)\|)]-0.5\}$	$\alpha = 20$, $\beta = 0.01$.
$l_{2}$-VSS-LMS [29]	$\mu (n)=\beta[1-\exp (-\alpha \|e(n)\|\times \\| \mathit{\boldsymbol{P}}(n)\\|^2)] \\ \mathit{\boldsymbol{P}}(n)=a\mathit{\boldsymbol{P}}(n-1)+be(n)\mathit{\boldsymbol{x}}(n)$	$\alpha = 1$, $\beta=0.001$, $a = 0.98$, and $b = 0.3$.
Proposed algorithm		$K = 4$. $t = 0.5$ for a white input and $t = 0.35$ for an AR(1) process.

New normalized LMS adaptive filter with a variable regularization factor

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 3

参考文献 33

相关文章 0

编辑推荐

Metrics

本文评价