基于递归对角神经网络自适应均衡算法的研究

Research on Adaptive Equalization Algorithm Based on Diagonal Recurrent Neural Network

阐述了神经网络自适应均衡算法的基本原理,提出了一种基于递归对角神经网络的自适应均衡算法,推导了算法的迭代公式,进行了计算机仿真。结果表明,随着传输信号信噪比的增加,稳态剩余误差减小,误码率降低。

This paper expatiates on the principle of adaptive neural network equalization algorithm. An adaptive equalization algorithm based on diagonal recurrent neural networks （DRNN） is proposed. Iteration formula is reduced and the algorithm is simulated by Matlab. The simulation shows that residual error and BER are reduced along with augment of SNR.