复数RBF神经网络自适应均衡算法研究  

Study on New Adaptive Complex Equalizers Constructed by Radial Basis Function Neural Networks

作　　者：王军锋[1] 张彬[1] 宋国乡[1] 

机构地区：[1]西安电子科技大学理学院数学系,陕西西安710071

出　　处：《系统工程与电子技术》2003年第7期848-850,859,共4页Systems Engineering and Electronics

基　　金：陕西省自然科学基金资助课题 ( 2 0 0 0SL0 2 )

摘　　要：在研究基于实数径向基函数 (RBF)神经网络均衡器结构的基础上 ,提出了几种新的适用于QAM信号的复数RBF神经网络自适应均衡器结构 ,并给出了相应的自适应算法。新的均衡器是充分利用了所得到的信号信息及RBF的特性而分别构成的。理论分析和计算机仿真结果都表明 。Several new complex equalizers applicable to QAM signals are presented. The equalizers, fully incorporating the obtained signal information and RBF's features, are constructed by radial basis function neural networks (RBFNs). The algorithms for them are also given. Theoretical analysis and experimental results show that the new algorithms have a better convergence property than those of the complex RBFNs based equalizers available.\;

关 键 词：神经网络 径向基函数 复数自适应均衡器 

分 类 号：TN911[电子电信—通信与信息系统]