摘要
在研究基于径向基函数(ＲＢＦ)神经网络的均衡器结构以及传统自适应均衡算法的基础上,提出了两种新的基于ＲＢＦ神经网络的自适应均衡器,并给出了相应的自适应均衡算法。新的均衡器是将判决反馈引入到ＲＢＦ神经网络中以及将Ａｄａｌｉｎｅ网络与ＲＢＦ网络有机的结合而分别构成的,仿真结果表明这两种新算法比基于ＲＢＦ神经网络的自适应均衡算法都具有更好的收敛性能。
Two kinds of equalizers based on RBF neural networks (RBFNs) are presented. The structure and algorithms of them are also given. The new equalizers are constructed by RBFNs with decision feedback and combined Adaline-RBFNs respectively. The experimental results show that the new algorithms have not only better convergence property but also less stable mean square errors than RBFNs-based equalizer.
出处
《信号处理》
CSCD
2002年第3期199-201,共3页
Journal of Signal Processing
基金
国家自然科学基金(69972034)资助
