基于Hopfield神经网络的盲均衡算法

Blind Equalization using Hopfield-type Neural Network algorithms

基于Hopfield神经网络的盲均衡算法是把发送信号盲恢复问题转换为求解带整数约束的二次规划问题,再利用hopfield网络来求解这个NP难的组合优化问题,从而直接恢复出信号。Hopfield神经网络分为两种,一种是连续空间的hopfleld神经网络,一种是离散空间的hopfield网络,文章介绍怎样用这两种方法对含公零点信道上发送的信号进行准确地恢复,基于Hopfield神经网络的盲均衡算法收敛速度快,性能明显优于其它经典算法。

we change the question of blindly recovering the transmitted signals into the one of quadratic optimization with integer constraints,in order to directly recover signals,Hopfleld networks can be devided into two kinds,One is distrete,The other is continuous.This paper introduces how to use the two methods to recover the signals transmitted on the common zero channels, the algorithm of this paper has the high speed of convergence and better performance than that of the traditional algorithm we mention in this Paper.