一种变步长NLMS-OCF算法的研究

Research on a variable step-size Normalized Least Mean Square algorithm with orthogonal correction factors

为了提高自适应滤波器的性能,提出了一种变步长的NLMS-OCF算法,直接将投影误差范数通过指数函数的映射得到平滑因子,进而对步长进行控制。仿真结果表明,相对于传统的NLMS-OCF算法,改进的算法获得了更好的收敛速度和稳态误差。

To improve the performance of the adaptive filters, this paper presents a variable step-size normalized least mean square algorithm with orthogonal correction factors, which uses an exponential function to obtain proper smoothing factors from the projected error norm. Simulation results show that the proposed algorithm is superior in terms of convergence rate and estimation error compared with traditional NLMS-OCF algorithm.