一种改进变步长LMS算法及其在系统辨识中的应用

被引:20
作者
李竹
杨培林
行小帅
机构
[1] 山西师范大学物理与信息工程学院
关键词
LMS算法; 变步长; 遗忘因子; 系统辨识;
D O I
10.19650/j.cnki.cjsi.2007.07.036
中图分类号
TN713 [滤波技术、滤波器];
学科分类号
摘要
在分析传统LMS算法及变步长LMS的基础上,本文提出了一种改进的变步长LMS算法并应用于系统辨识。新算法利用瞬时误差的四次方和遗忘因子共同来调整步长,进一步解决了收敛时间和稳态误差的矛盾。将新算法应用到系统辨识中时,与传统LMS算法和变步长LMS算法相比,仿真表明新算法有更快的收敛速率。
引用
收藏
页码:1340 / 1344
页数:5
相关论文
共 11 条
  • [1] A novel kurtosis driven variable step - size adaptive algorilhm. D I Pazaitis,,A G Constantinidrs. IEEE Transactions on Signal Processing . 1999
  • [2] Variable Step - Size NLMS and Affine Projection Algorithms. Hyun - Chool Shin,,Ali H Sayed. IEEE SIGNAL PROCESSING LETTERS . 2004
  • [3] A class of adaptive step-size control algorithms for adaptive filters. KOIKE S. IEEE Transactions on Signal Processing . 2002
  • [4] Stochastic gradient adaptive filters with gradient adaptive step-size. MATHEWS V J. Proc.IEEE Int.Conf. Acoust.Speech,Signal Processing . 1990
  • [5] On convergence and selecting step size for the LMS algorithm. BINNAN P. Journal of China Institute of Communications . 1994
  • [6] A robust variable step-size LMS- type algorithm:analysis and simulations. TYSEER A,MAYYAS K. IEEE Transactions on Signal Processing . 1997
  • [7] The effects of large interfer- ence on the tracking capability implemented echo cancellers. GITLIN R D,WEINSTEIN S D. IEEE Transactions on Computers . 1978
  • [8] Stationary and no stationary learning characteristics of the LMS adaptive filter. WIDROW B,MCCOOL J M,ARIMORE M G. Proceedings of Tricomm . 1976
  • [9] Adaptive tittering theroy. HAYKIN S. . 1986
  • [10] Adaptive filtering with averaging. YIN G. Adap- tive Control,Filtering,and Signal Processing . 1995