SUN Yang,CHEN Guang-tuo,DU Yu-Ming,et al.A Study of Variable Step Size LMS Algorithm based on Hyperbolic Cosine[J].Journal of Chengdu University of Information Technology,2017,(06):605-608.[doi:10.16836/j.cnki.jcuit.2017.06.006]
一种基于双曲余弦型变步长LMS算法研究
- Title:
- A Study of Variable Step Size LMS Algorithm based on Hyperbolic Cosine
- 文章编号:
- 2096-1618(2017)06-0605-04
- 分类号:
- TN911.4
- 文献标志码:
- A
- 摘要:
- 变步长LMS算法较好的克服了固定步长LMS算法中收敛速度与稳态误差之间不可兼得的矛盾。变步长LMS算法中步长因子μ应该按照一定的要求变化,即在误差较大时取较大的μ值,误差较小时取较小的μ值,直到随着误差接近于零,μ值也缓慢收敛于零值。建立了步长因子与误差信号之间的一种双曲余弦非线性函数模型,在此模型的基础上提出一种新的LMS算法,将该算法对比传统固定步长LMS算法和Sigmoid型变步长LMS算法仿真分析,仿真结果表明,新的LMS算法在不失精度和计算复杂度的情况下,算法性能较好。
- Abstract:
- The variable Step Length Factor LMS algorithm can overcome the contradiction between the convergence speed and the steady-state error in the fixed step size LMS algorithm. The step factor μ should change according to certain requirements in variable step size LMS algorithm. The step factor μ takes a larger value when the error is bigger or a smaller value when the error is smaller, and the value of The step factor μconverges slowly to zero while the error close to zero. This paper establishes a nonlinear hyperbolic cosine function model between the step size and the error signal.Based on this model, this paper proposes a new LMS algorithm. It made a simulation the new algorithm and conditional fixed step size LMS algorithm and Sigmoid's type variable step LMS algorithm. The simulation results show that the new LMS algorithm has a good performance without losing accuracy and computational complexity.
参考文献/References:
[1] 罗小东,贾振红,王强.一种新的变步长LMS自适应滤波算法[J].电子学报,2006,34(6):1123.
[2] B Widrow,M E Hoff.Adaptiveswitchingcircuits[J].Ire Wescon Conv Rec,1966,4.
[3] 西蒙赫金.自适应滤波器原理(第四版)[M].北京:电子工业出版社,2003.
[4] 汪成曦,刘以安.改进的最小均方自适应滤波算法[J].计算机应用,2012,32(7).
[5] 覃景繁,韦岗.基于S型函数的变步长LMS自适应滤波算法[J].无线电工程,1996,26(4):44-47.
[6] 王敏强,郑宝玉.一种新的可变步长自适应滤波算法[J].信号处理,2004,12(6).
[7] 高鹰,谢胜利.一种变步长LMS自适应滤波算法及分析[J].电子学报,2001,29(8):1094-1097.
[8] 刘晓志,吴永刚.基于双曲余弦函数的智能天线自适应波束形成算法[J].新型工业化,2014,4(3):74-79.
[9] 黄甫堪,陈建文,楼生强.现代数字信号处理[M].北京:电子工业出版社,2003.
备注/Memo
收稿日期:2017-09-11 基金项目:国家自然科学基金资助项目(41475143)