一种变步长的零吸引归一化自适应滤波算法  

A Zero Attraction Normalization of Variable Step Size Adaptive Filtering Algorithm

作　　者：火元莲[1] 徐天赐 齐永锋 徐玉荣 柳洁 HUO Yuanlian;XU Tianci;QI Yongfeng;XU Yurong;LIU Jie(College of Physics and Electronic Engineering,Northwest Normal University,Lanzhou Gansu 730070,China;College of Computer Science and Engineering,Northwest Normal University,Lanzhou Gansu 730070,China)

机构地区：[1]西北师范大学物理与电子工程学院,甘肃兰州730070 [2]西北师范大学计算机科学与工程学院,甘肃兰州730070

出　　处：《传感技术学报》2024年第7期1216-1222,共7页Chinese Journal of Sensors and Actuators

基　　金：甘肃省自然科学基金项目(23JRRA692);国家自然科学基金项目(62267007)。

摘　　要：为了使最小均方算法对高斯噪声环境和稀疏系统具有更好的收敛速度和稳态误差,提出了一种变步长的零吸引归一化最小均方算法。该算法将改进的Versoria函数的稀疏感知范数与归一化最小均方算法相结合后,引入了一种新的类高斯函数的变步长策略,解决了固定步长条件下算法收敛速度较慢、跟踪性能较差的问题。从理论层面分析了所提算法的收敛性,并基于MATLAB平台讨论了改进的类高斯步长函数中各参数对算法性能的影响。最后将所提算法与其他同类算法应用于不同信噪比条件下的高斯噪声环境以及稀疏环境中进行未知系统辨识实验,仿真结果表明,所提算法具有更快的收敛速度、更好的跟踪能力以及较小的稳态误差。In order to improve the convergence speed and steady-state error of the least mean square algorithm for gaussian noise environment and sparse system,a variable step size zero attraction normalized least mean square algorithm is proposed,which combines the improved sparse perception norm of Versoria function with the normalized least mean square algorithm,and introduces a new gaussianlike variable step size strategy,solving the problems of slow convergence and poor tracking performance under the condition of fixed step size.The convergence of the proposed algorithm is analyzed theoretically,and the influence of parameters in the improved Gaussian-like step function on the performance of the algorithm is discussed based on MATLAB platform.Finally,the proposed algorithm and other similar algorithms are applied to unknown system identification experiments in gaussian noise environment and sparse environment with different SNR conditions.Simulation results show that the proposed algorithm has faster convergence speed,better tracking ability and smaller steady-state error.

关 键 词：自适应滤波 最小均方算法 归一化 类高斯函数 零吸引 

分 类 号：TN713[电子电信—电路与系统]