利用高阶累积量和独立分量分析网络进行盲均衡与系统辨识

HOS and ICA Based on Blind Equalization and System Identification

在高阶累积量的基础上，利用过采样技术和独立分量分析神经网络，得到一种新的自适应盲辨识和信道均衡方法。与同类方法相比，本文提出的方法网络结构简单，不必利用训练序列，可以同时得到系统参数和均衡输出。

A new algorithm for blind identification and equalization is proposed. Based on higher order cumulant, the proposed algorithm combines over sampling technique with ICA neural networks. Compared with the existing algorithms, the proposed one does not use learning sequence, has a simple architecture and can give the channel reconstruction and signal reconstruction simultaneously.