LMS算法自适应噪声抵消系统参数的最佳设计

威德罗(Widrow)提出的LMS算法自适应噪声抵消系统中,由于d(j)=s(j)+n_0,带有随机噪声n_0,不再是一个理想的期望响应。这时自适应滤波器除了有由于LMS近似算法造成的权失调外,还有随机噪声的影响。本文讨论了这种随机噪声对系统性能的影响,以及如何选择系统的加权数目n,收敛因子μ这两个主要参数,使系统性能达到最佳。最后,对得出的结论进行了实验验证。

Optimal Parameter Design of Adaptive Noise Cancelling by LMS Algorithm

The performance and major parameter of adaptive filter acquired by using Widrow LMS algorithm is limited at the condition of s(j) as an ideal response. But, in adaptive noise cancelling system, using LMS algorithm d(j)-s(j)+n0 , with noise n0, is no longer an ideal expect response. There is a weight overshoot causing by LMS algorithm and an error causing by noise n0 in adaptive filter. In this paper, at first the error caused by noise n0 is discussed, then how to select the optimal parameters to obtain the best performance, finally the result is tested.