一种变步长因子及变动量因子的自适应算法

An Adaptive Algorithm with a Variable Step and a Variable Momentum Factor

作为对动量ＬＭＳ算法的补充和改进，提出了一种带有变步长和变动量因子的自适应算法。该算法将传统的动量ＬＭＳ算法中的常量步长调节因子与常系数的动量调节因子均改变成随机变量，从而为自适应滤波器提供了更好的收敛特性。计算机模拟的结果表明，给出的算法在平滑随机扰动和加快收敛速度方面起到了一定的作用。

As a generalization of the momentum LMS algorithm,an adaptive algorithm witha variable step and a variable momemtum factor is proposed.The adaptive algorithm changesthe constant step and the constant momentum factor in conventional MLMS algorithm to astochastic variable so as to provide a better convergence property for adaptive filter,whichhas been supported by computer simulations.