A New Approach for on-Line Blind Equalization and Channel Identification

本文提出一种新的盲均衡和通道辨识方法．该方法基于过采样技术和独立分量分析神经网络进行信道均衡，利用均衡输出序列和高阶累积量，辨识信道参数．与通常的方法相比，该方法网络结构简单，不必利用训练序列．

A new approach for blind equalization and channel identification is proposed in this paper. The equalization scheme is based on over sampling technique and an independent component analysis network. The equalized sequence and its higher order statistics are used to identify the channel parameters. Compared to traditional equalization methods, the proposed approach is with a simple architecture, and does not need learning sequences. Computer simulations show the validity of the proposed method.