RLS算法的改进性能分析(英文)  

Improved Performance Analysis of Recursive Least Squares Algorithm

作　　者：熊鹰[1] 梁树雄[2] 尹俊勋[1] 

机构地区：[1]华南理工大学电子与信息工程学院,广东广州510640 [2]香港城市大学电子工程系

出　　处：《华南理工大学学报（自然科学版）》2001年第11期32-36,共5页Journal of South China University of Technology(Natural Science Edition)

摘　　要：讨论了递归最小均方误差算法 (RLS)的性能分析 .对该算法误差的均值及方差都进行了研究 .在平均原理基础上使用了更精确的处理方法 ,得到一种改进的性能分析方法 .新的分析方法得到的理论值更接近模拟的结果 ,优于通常方法的结果 .This paper discussed the performance analysis of the recursive least squares (RLS) algorithm. Both mean and mean square behavior are studied. By using a novel and more accurate approaching method, accompany with averaging principle, an improved performance analysis of mean square behavior is carried out. It is shown that the theoretical analysis and the simulation result are close to each other, outperforming of the conventional analysis of RLS algorithm. The stability of the basic RLS algorithm is discussed under the new analysis result.

关 键 词：递归最小均方误差算法 均方误差分析 稳定性 性能分析 均值 方差 平均原理 

分 类 号：TN911.7[电子电信—通信与信息系统]