基于Sigmoid函数的变步长LMS自适应滤波算法性能分析

Performance Analysis of LMS Adaptive Filtering Algorithm with Variable Step Based on Sigmoid Function

总结了最小均方(LMS)、基于Sigmoid函数的变步长(SVS)-LMS、改进的SVS-LMS和基于误差相关的变步长LMS自适应滤波算法,讨论了各算法的收敛性能、跟踪性能和稳态误差,并通过计算机仿真,分析、验证各种变步长算法的不同性能表现以及误差阶数对算法性能的影响,给出了合理的建议。

This paper summarizes the adaptive filtering algorithms of least mean square （LMS）,var- iable step size （SVS）-LMS based on Sigmoid funtion,modified SVS-LMS,variable step LMS based on error correlation, discusses the convergence performance, tracking performance and steady errors of each algorithm, analyzes and validates the different performances of each adaptive variable step algorithm through computer simulation and the influence of error order on the algorithm perform- ance, presents the reasonable suggestions.