基于FastICA的盲源分离算法研究

Fast ICA based Blind Source Signal Separation

FICA是一种非常有效的盲信号分离算法,该算法比批处理甚至自适应处理具有更快的收敛速度。本文介绍了一种以四阶累计量为判据的固定点算法。运用matlab仿真,对四路不相关的源信号进行分离,仿真结果表明,该算法能有效的分离出四路源信号,获得了比较好的效果。

The fast independent component analysis is a effective algorithm used in blind source separation,this algorithm have the fastest convergence speed more than the batch processing and the adaptive processing. This paper introduces affixer-point algorithm which uses fourth order cumulant as the optimized criterion. Four source signals were discreted by simulation using Madab,the results showed that this algorithm can discrete four signals effectively.