摘要
通过对传统最小均方误差(leastmean square,LMS)算法迭代因子μ进行分析,讨论了μ与收敛速度及稳态失调的关系,在此基础上研究了一种新的变步长LMS自适应算法,建立了步长因子μ与输入信号及迭代次数n之间的一种新的非线性关系。通过理论分析,该算法与传统LMS算法相比,其收敛速度更快、稳态误差较小,且计算量增加不大,采用Matlab仿真表明了该算法的优越性。
Traditional LMS algorithm of step size factor was discussed and a new variable step size LMS auto-adapted algorithm was proposed.A new nonlinear functional relationship among the step size factor,iteration number of times and the input signal was established.Compared with the traditional LMS algorithm,this algorithm has faster convergence speed and minor steady state error while computational complexity is not increased too much.The Matlab simulation results showed the superiority of the algorithm.
出处
《武汉理工大学学报（信息与管理工程版）》
CAS
2010年第4期547-549,553,共4页
Journal of Wuhan University of Technology：Information & Management Engineering
关键词
变步长LMS算法
收敛速度
稳态失调
噪声抵消
variable step size LMS algorithm
convergence speed
steady state maladjustment
de-noising
