一种新的变步长LMS自适应算法及其应用

传统的固定步长的LMS算法难于同时获取较快的收敛速度与较小的稳态误差,基于这一矛盾,本文在分析基本LMS算法的基础上,提出了一种新的变步长LMS算法,该算法具有较快的收敛速度和较小的稳态误差。理论分析和计算机仿真结果表明该算法在收敛速度与稳态误差的性能上均优于基本LMS算法。尤其是在低估噪比的情况下,其性能的优越性更为突出。与传统的LMS算法相比较,新算法更适合于应用到回波消除等实时性要求高的场合。

New variable step-size LMS algorithm and its simulations

The conventional fixed step size LMS algorithm is difficult to exquisite the fast convergence speed and low steady state error.On the basis of the problem,a new LMS a lgorithm of variable step size is proposed based on a brief discussion on standard LMS.The new algorithm has higher speed of convergence than the standard LMS algorithm.The performance of this paper algorithm is better than that of standard LMS with the theoretical analysis and computer simulations,especially in low SNR situation.The new algorithm is much better when applied to echo elimination.