一种变正则化矩阵的改进多带结构子带自适应滤波算法

摘要

定正则化因子的改进多带结构子带自适应滤波(IMSAF)算法在取得收敛速度快和稳态失调误差小之间存在冲突.根据系统噪声抵消原理,设定子带后验误差功率等于子带噪声功率,本文提出了变正则化矩阵的IMSAF算法来解决这一问题.仿真结果证明,所提算法可以同时达到收敛速度快、稳态失调误差小以及追踪速度快等优势.%The conflict between fast convergence rate and small steady-state misadjustment is a problem of the improved multiband-structured subband adaptive filter (IMSAF) algorithm with a fixed regularization factor.Based on the principle of minimizing the system noise,the power of the subband posteriori error is assumed to equal the power of the subband noise.Under this assumption a variable regularization matrix for the IMSAF algorithm is proposed to solve this problem.Simulation results show that the proposed method is superior to the IMSAF algorithm in terms of convergence rate,steady-state misadjustment and tracking capability.