变步长NLMS自适应滤波算法研究

ON ADAPTIVE FILTERING ALGORITHM FOR VARIABLE STEP-SIZE NLMS

将步长因子μ与误差信号e(n)之间的一种非线性函数关系引入归一化最小均方(NLMS)自适应滤波器,提出了一种变步长NLMS算法。与一些已有算法相比,算法的步长因子易于设计和控制,对于有色输入信号,能够得到较快的收敛速度和较低的稳态失调。仿真结果和理论分析相一致,证实了算法的有效性。

By introducing a nonlinear functional relationship between step size parameter μ and error signal e （n） into Normalized LeastMean-Square （NLMS） adaptive filter, a variable step-size NLMS algorithm is proposed. Compared with some existing approaches, the proposed step-size parameter of the algorithm in this paper is easily designed and controlled, and can obtain fast convergence rate and low misadjustment for coloured input signal. Simulation results confirm the theoretical analysis and verify the efficacy of the new approach.