一种新的步长等变化自适应算法

定义了描述信号分离状态的一种测度，并在认真分析相关固定步长和变步长EASI算法的基础上。提出了一种新的步长自适应等变化自适应（EASI）算法。该算法步长是基于分离状态的。其学习速率由信号的分离程度自适应地选取。因而能很好地解决收敛速度和稳态误差之间的矛盾。计算机仿真结果与理论分析相一致，证实了该算法明显优于传统的EASI算法．

A New Step-size EASI Algorithm

This paper definites a measurement method to describe the signal separation state. After detailed analyzing relevant fixed step-size and variable step-size EASI algorithms, a new self-adaptive step-size EASI algorithm is presented. The step-size of this algorithm is based on separation state,its learning ratio is chosen adaptively according to separating degree, therefore it can resolve the contradiction between the convergence speed and misadjustment error. Computer simulation result and the theoretical analysis is identical, which confirms the algorithm is superior to traditional EASI algorithms.