基于相对误差互相关函数的变步长LMS算法

许多时变步长自适应滤波算法被用来解决标准LMS算法的固有矛盾,但实验表明这些算法易受独立噪声的干扰,或计算量大、耗时太长。针对上述问题,提出了一种新的变步长LMS自适应滤波算法,它采用相对的误差互相关函数来控制步长更新。该算法计算量小、易于控制,具有快速的收敛速度和较小的失调,不受已存在的非相关噪声的影响,可很好地应用于自适应对消系统中,且在低信噪比环境中仍能保持良好的性能。计算机仿真及实测数据的处理与理论分析结果一致。

Variable step-size LMS algorithm based on cross-correlation of relative output error signals

Many time-varying step-size adaptive algorithms have been proposed to resolve the inherent contravention of the conventional LMS algorithm.But experiments of those algorithms indicate that their performance is highly sensitive to the noise disturbance or they are too time-consuming.A new variable step-size LMS adaptive filtering algorithm is presented,which makes use of the intercorrelation function of the relative output error signal to control the new variable step-size.Due to its smaller computus,faster convergence rate and smaller misadjustment,and its performance is not influenced by existing uncorrelated noise,the proposed algorithm can be applied perfectly in the adaptive noise cancellation system and achieves favorable performance in lower SNR situation.Results of computer simulation and practical signal processing are coincident with the theoretical analyses.