一种新的可变步长LMS算法及其性能分析  

Performance analysis of a novel variable step size LMS algorithm

作　　者：王伟强[1] 杨金明[1] 杨苹[1] 

机构地区：[1]华南理工大学电力学院,广州510640

出　　处：《信息技术》2005年第9期117-119,122,共4页Information Technology

基　　金：广东省自然科学基金(020906)

摘　　要：为了解决基本LMS(Least Mean Square)算法中收敛速度和稳态误差之间的矛盾,提出了一种新的可变步长LMS算法(λNVSS)。该算法通过引入修正系数β和遗忘因子λ,利用当前和过去共M(滤波器阶数)个误差信息来决定下一步的迭代步长。仿真结果显示,λNVSS算法对于平稳过程中的滤波器,能获得较NVSS更快的收敛速度和更小的稳态误差。同时,λNVSS算法还有较好的跟踪跃变系统的能力。In this paper, a novel variable step size LMS （Least Mean Square） algorithm （λNVSS） with error nonlinearities is provided to enhance the trade - off between misacljustment and speed of convergence. A modified coefficient（β） and a forgetting factor（λ） are introduced in the algorithm. The step size of the new algorithm is inversely with the summation of squared norm of the error vector and β/λ. Simulation results demonstrated that the new algorithm gets significant improvements in the speed of convergence with a very small value of misadjustment in stationary environments and better tracking capability.

关 键 词：LMS算法 自适应滤波 NLMS NVSS λNVSS 遗忘因子 

分 类 号：TN911[电子电信—通信与信息系统]