偏差补偿稀疏自适应滤波算法

Bias-Compensated Sparse Adaptive Filtering Algorithm

为解决输入信号受到噪声干扰和系统发生时变导致ZA-LMS算法估计精度下降的问题,提出一种偏差补偿稀疏自适应滤波算法。算法是在LMS算法的代价函数中添加了l_(1)范数,并导出新的滤波器权系数更新公式,使该算法获得一个指向零矢量的修正量。将无偏准则应用于算法的自适应迭代公式,推导出的偏差补偿项,用以修正输入噪声带来的偏差。仿真实验结果表明,算法在输入信号受噪声干扰和系统发生时变时,具有良好的收敛性、稳态性和跟踪性。

To solve the problem that the estimation accuracy of ZA-LMS algorithm decreases due to noise interference of input signal and system time-varying,a deviation compensation sparse adaptive filtering algorithm is proposed.The algorithm adds l_(1) norm to the cost function of LMS algorithm,and derives a new updating formula of filter weight coefficient,so that the algorithm can obtain a correction value pointing to zero vector.Unbiased criterion is applied to the adaptive iterative formula of the algorithm,and the deviation compensation term is derived to correct the deviation caused by input noise.Experimental results of simulation show that the algorithm has good convergence,stability and tracking when the input signal is disturbed by noise and the system is time-varying.