宋國棟 張懷遠

摘要：通過對歸一化LMS算法（NLMS）進行拉格朗日方程最小值運算而得到的CS-LMS算法具有良好的收斂性和靈活性。該文為進一步解決自適應濾波算法不能有效處理既要求收斂速度快又要求穩(wěn)態(tài)誤差小的矛盾,建立了步長與信號誤差之間的一種非線性函數(shù)模型，并將改進的CS-LMS算法應用到混沌通信中。仿真結果表明,改進算法的收斂速度和穩(wěn)態(tài)誤差性能都有較大的提高。

關鍵詞：LMS；步長；非線性函數(shù)模型；自適應CS-LMS算法；混沌通信

中圖分類號：TP311文獻標識碼：A文章編號：1009-3044(2012)16-3867-02

An Improved Adaptive CS-LMS Algorithm and its Performance Analysis in Chaotic Communication

SONG Guo-dong, ZHANG Huai-yuan

(Electronic and Information Engineering College of Southwest University , Chongqing 400715,China)

Abstract: CS-LMS algorithm provides faster convergence and higher flexibility than the normalized least mean square (NLMS) by mini? mizing the Lagrangian function. This paper, in order to solve adaptive filters conflict of gaining the fast convergence speed and low steady state error, a non- linear functional model between step size and signal error will be established, and we will continue to apply this new al? gorithm to chaotic communication. Simulation results present the proposed algorithm enhances the speed of convergence and quality of sta? bility distinctly.

Key words: LMS; step size; a non- Linear functional model; adaptive CS-LMS filter; chaotic communication

該文給出的改進的CS-LMS算法,是在步長參數(shù)μ與誤差信號e(n)之間建立了一種新的非線性函數(shù)關系,該算法與傳統(tǒng)LMS和CS-LMS算法自適應濾波算法相比，具有較快的收斂速率并且在高信噪比下穩(wěn)態(tài)誤差方面表現(xiàn)的比較優(yōu)越，但在仿真中也發(fā)現(xiàn)，非線性函數(shù)中的參數(shù)對算法性能有著決定性的影響，所以一定要結合實際情況合理選擇各類步長參數(shù)。

[1]王秀芳.基于改進LMS算法的自適應濾波器[J].大慶石油學院學報,2009,33(6).

[2]余榮貴,葉水生.一種改進的變步LMS自適應濾波算法及其在噪聲抵消中的應用[J].計算機工程與科學,2009 ,31 (4) :66 - 68.

[3] Huang H, Lee J.A New Variable Step-Size NLMS Algorithm and Its Performance Analysis[J].IEEE Tran. Signal Process., doi: 10.1109/ TSP.2011.2181505, 2011.

[4]曾偉. LMS自適應濾波算法的改進及性能分析[J] .計算機仿真,2011, 28(4):111-114.

[5]高鷹,謝勝利.一種變步長LMS自適應濾波算法及分析[J].電子學報,2001(8).

[6]徐凱.一種改進的變步長自適應濾波器LMS算法[J].電路與系統(tǒng)學報,2001 ,9(4).

[7] Leung H, Lam J,Sect S R,et al.Design of demodulator for the chaotic modulation communication system[J].IEEE Tran Circuits Sys I, 1997, 44(3):262-267.