基于Rayleigh分布曲线的混合权系数变步长LMS算法  

Variable step-size LMS algorithm with hybrid weight coefficients based on the Rayleigh distribution curve

作　　者：陈东伟 刘卫东 张明怡 金梦哲 方庆园 CHEN Dongwei;LIU Weidong;ZHANG Mingyi;JIN Mengzhe;FANG Qingyuan(School of Traffic and Transportation,Shijiazhuang Tiedao University,Shijiazhuang 050043,China;Hebei Provincial Key Laboratory of Electromagnetic Environmental Effects and Information Processing,Shijiazhuang 050043,China)

机构地区：[1]石家庄铁道大学交通运输学院,河北石家庄050043 [2]河北省电磁环境效应与信息处理重点实验室,河北石家庄050043

出　　处：《通信学报》2025年第3期62-73,共12页Journal on Communications

基　　金：国家自然科学基金资助项目(No.61801309,No.51807123);国家重点实验室开放课题基金资助项目(No.JCKYS2022DC07)。

摘　　要：针对已有变步长LMS自适应滤波算法无法兼顾收敛速度、跟踪性能和稳态误差的局限性,提出了一种基于Rayleigh分布曲线的混合权系数变步长LMS算法。详细讨论了新的变步长函数中参数α、β、γ和混合滤波权系数W(n)对所提算法性能的影响以及抗干扰性分析。仿真与实测结果表明,所提算法在减少计算复杂度的同时,将收敛速度提升了至少12%,且在不同信噪比下均展现出良好的收敛速度、跟踪性能和稳态误差。所提算法适用于复杂环境中的电磁干扰消除。Due to the limitations of existing variable step-size LMS adaptive filtering algorithms in addressing convergence speed,tracking performance,and steady-state errors,a variable step-size LMS algorithm with hybrid weight coefficients based on the Rayleigh distribution curve was proposed.The impact of the new variable step function parametersα,β,γand the hybrid filter weight coefficient W(n)on the proposed algorithm performance and interference suppression analysis were discussed in detail.Simulation and experimental results demonstrate that the proposed algorithm reduces computational complexity while improving convergence speed by at least 12%.Moreover,it exhibits superior convergence speed,tracking performance,and steady-state errors across various signal-to-noise ratios.The proposed algorithm is suitable for electromagnetic interference cancellation in complex environments.

关 键 词：自适应滤波 最小均方算法 Rayleigh分布曲线 混合权系数 电磁干扰 

分 类 号：TN927[电子电信—通信与信息系统]