基于Volterra级数扩展模型的变步长VFxBSLMS算法  

Variable step size VFxBSLMS algorithm based on Volterra series expansion model

作　　者：仝喜峰 陈卫松[1] 钱隆彦 TONG Xifeng;CHEN Weisong;QIAN Longyan(College of Physics and Eelectronic Information,Anhui Normal University,Wuhu 241000,China)

机构地区：[1]安徽师范大学物理与电子信息学院

出　　处：《传感器与微系统》2019年第10期135-138,共4页Transducer and Microsystem Technologies

基　　金：国家自然科学基金青年科学基金资助项目(61601005);安徽省自然科学基金资助项目(1608085QF138)

摘　　要：针对脉冲噪声的控制,提出了一种基于Volterra级数扩展模型的变步长VFxBSLMS算法。通过增加误差信号预处理环节,采用非线性压缩的方式,有效地解决了冲击样本对算法稳定性的影响。针对输入信号的强相关性及算法收敛速度问题,将变步长算法思想引入到非线性滤波中。仿真结果表明:VFxBSLMS算法在不同强度脉冲噪声背景下,都能达到较好的收敛效果。Aiming at the control of impulse noise,a variable step size VFxBSLMS algorithm based on Volterra series expansion model is proposed.The algorithm effectively solves the influence of impact samples on the stability of the algorithm by adding the error signal preprocessing part and nonlinear compression method.Aiming at the problems of the strong correlation of input signals and the convergence speed of the algorithm,the variable step size algorithm idea is introduced into the nonlinear filtering.Simulation results show that the VFxBSLMS algorithm can achieve better convergence under impulse noise background with different intensity.

关 键 词：S型函数 VOLTERRA滤波器 稳态分布 脉冲噪声 

分 类 号：TB535[理学—物理]