具有周期输入的频率域自适应滤波器的健壮性

自适应滤波在工程技术中有着极其广泛的应用,例如,用于消除回声、横向均衡、降低噪声等.虽然存在多种类型的自适应滤波算法,但在实际应用中最经常采用的是 Widrow算法.这是因为这种算法虽非最优但很容易实现,且具有较好的性能.时间域中的Widrow 算法可以简述如下:设{x_k}和{y_k}分别是输入序列和理想输出序列.在 k 时刻作估计(?)_k=sum from n=0 to N-1 h_n~((k))x_(k-n),相应的误差为 e_k=(?)_k-y_k.Widrow 自适应滤波器是指递推算法

ON THE ROBUSTNESS OF THE ADAPTIVE FILTER IN FREQUENCY DOMAIN WITH PERIODIC INPUT

In this paper we study the robustness of the Widrow filter in frequency domain withperiodic input in the noiseless case.The convergence and divergence conditions for theWidrow filter are presented when the correlative input appears in a certain term of the ite-ration instead of the real input.The results imply that the phases play an important rolein the robustness problem of the adaptive filter in frequency domain.