GACI算法权系数的时间相关特性

Time Correlation Statistics of the Generalized Adaptive Coherent Integrated Algorithm Weights

GACI算法是在ACI算法基础上提出来的,是一种新的自适应信号处理算法。本文采用特征值特征向量分析法,对其权系数的时间相关特性进行了分析,给出了在非相关高斯数据条件下,任意时刻权系数向量相关矩阵的递推关系式,同时得到了在稳态情况下,相关矩阵的更简捷的关系式,并证明了使输入数据相关矩阵R_(XX)r对角化的模态矩阵Q,同样使权系数误差向量V的相关矩阵R_(XX)r也对角化,从而权系数相关矩阵的求解问题转化为线性方程组的求解问题。借此可以对算法的性能作进一步的研究探讨。

The Least Mean Square (LMS) algorithm has widely been applied because of its simplicity, but its performance is often unsatisfactory[1-4]. Adaptive Coherent Integrated (ACI) algorithm is the newest modification of LMS algorithm, improving LMS through the introduction of an integrated term into the weight updating equation [1]. Starting from ACI algorithm, the author in a previous paper[5] proposed GACI (Generalized Adaptive Coherent Integrated)algorithm which is better than ACI. In GACI more integrated term...