基于离散Hopfield神经网络的盲信号均衡新算法

A new blind equalization algorithm based on discrete Hopfield-type neural network

针对许多现有的盲均衡算法对发送信号,均衡器,信道等都有不同的要求,本文在利用接收数据正交补投影和发送信号属于有限字符集的先验知识之上,构造出直接待恢复盲信号的二次规划问题,推出了一种基于离散Hopfield神经网络直接盲信号均衡的新方法,该算法有效地克服了上述问题,在SIMO(单输入多输出)信道上的仿真结果表明:该算法收敛速度快,性能好。

Focused on many existing blind equalization algorithms that have different requests for sending signals, equalization machine, and channel,and based on a quadratic programming problem to blindly detect transmitted signals with the projection of the orthogonal complement of a received data matrix and a priori knowledge that transmitted signals belonged to finite alphabets, a new direct blind equalization using discrete hopfield - type neural network algorithm is proposed by the paper. This algorithm conquers above problems effectively. Based on SIMO channel, simulations show that the algorithm achieves good astringency and satisfactory performance.