一种新的神经网络均衡器:结构、算法与性能

A Novel Adaptive Neural Network Equalizer Structure Algorithm and Performance

本文根据克服数字通信中码间干扰（ISI）的最佳均衡解一般表达式，提出了一种新的自适应神经网络均衡器结构，然后导出了基于该结构的一种自适应算法和相应的学习规则，最后对提出的自适应神经网络均衡器性能进行了计算机模拟，模拟结果与分析表明：本文提出的神经网络均衡器用于实现最佳信道均衡非常有效，比传统线性均衡器和Gibson等人[1]提出的多层感知均衡器（MLPE）性能更优越，更具实用性．

This paper suggests a novel adaptive neural network channel equalizer to overcome the in ter-symbol interference (ISI) existing in Uarious digital Communication systems, based on the general optimum equlization solutions, derives its adaptive algorithms and learning rules, and in vestigates the performance of the novel neural network equalizer appied to some sommunication systems by means of compputer simulations. it is conchuded that the novel neural equalizer is very effective in implementing the optimum channel equalization and it has superior performances over both the traditional linear equalizers and the multilarye perceptron equalizers (MLPE) suggested by Gibson, et al.