摘要
在详细分析并吸收几种变步长最小均方差(LMS)算法优点的基础上,给出了一种通用形式的变步长LMS算法。采用遗传算法对该算法进行参数优化设计,能够快捷高效地得到α、β、m及h参数优化组合,而不需要通过以往经验或试凑等方法来获取各参数的最优值。Matlab仿真与理论分析相一致,表明变步长LMS算法具有良好的收敛性和跟踪性,可广泛地应用于数字通信系统中。
On the basis of analyzing and absorbing the advantages of several variable step size LMS algorithms, the variable step size LMS algorithm with general form is given. By using genetic algorithm, the parameters of the algorithm are optimized, thus the optimal combination of parameters α,β,m and h can be obtained quickly and efficiently, while it is not necessary to use traditional methods, e.g. , by experience or cut and try, to acquire the optimal value of each parameter. The consistency of the Matlab simulation and theoretical analysis indicates that the variable step size LMS algorithm possesses excellent convergence and traceability, thus it can be widely used in digital communication systems.
出处
《自动化仪表》
CAS
北大核心
2012年第12期17-20,共4页
Process Automation Instrumentation
基金
江西省自然科学基金资助项目(编号:CA201004007)
江西省科技支撑计划基金资助项目(编号:CB200904355)
关键词
遗传算法
LMS算法
收敛性
跟踪性
数字通信
参数优化
Genetic algorithm LMS algorithm Convergence Traceability Digital communication Parameters optimization
