摘要
为解决改进的基于Sigmoid函数变步长最小均方(G-SVSLMS)算法步长更新公式易受噪声干扰的问题,根据高斯白噪声的零均值特性,对G-SVSLMS算法进行改进,提出基于零均值特性的改进G-SVSLMS算法,使算法的抗噪声干扰能力明显增强。理论分析和仿真结果表明:若两算法选取相同参数α、β,则基于零均值特性的改进G-SVSLMS算法相对于G-SVSLMS算法具有小的稳态误差;在保证算法收敛的条件下,基于零均值特性的改进G-SVSLMS算法相对于G-SVSLMS算法具有较快的收敛速度。
The G-SVSLMS algorithm＇s step-formula can be disturbed easily by noise jamming. According to the characteristics that the mean of white Gaussian noise is zero, improved G-SVSLMS algorithm based on zero mean characteristics is put forward in order to improved G-SVSLMS algorithm＇s ability of anti-noise. If two algorithms choose the same parameters α、β, improved G-SVSLMS algorithm will have less steady-state error than G- SVSLMS algorithm. Under the condition that the two algorithms are convergent, the convergence rate of im- proved G-SVSLMS algorithm is bigger than that of G-SVSLMS algorithm. Improved G-SVSLMS algorithm＇s per- formance is testified through theoretical analysis and simulation.
出处
《电讯技术》
北大核心
2013年第3期284-287,共4页
Telecommunication Engineering
关键词
信号处理
噪声抑制
G—SVSLMS算法
零均值特性
稳态误差
收敛速度
signal processing
noise suppression
G-SVSLMS algorithm
characteristic of zero mean
steady-state error
convergence rate
