基于自适应投影学习算法的RBF神经网络多用户检测方法  

A Multiuser Detection Approach Using RBF Neural Networks based on Adaptive Projective Learning Algorithm

作　　者：王焱滨[1] 胡志恒[1] 虞厥邦[1] 

机构地区：[1]大电子科技大学光电子技术系,成都610054

出　　处：《信号处理》2002年第6期518-521.,共4页Journal of Signal Processing

摘　　要：自适应投影学习算法是一种简单有效的构造和训练径向基函数神经网络的方法,该方法能迭代地确定径向基函数的个数,中心的位置以及网络的权系数。本文将基于自适应投影学习算法的径向基函数神经网络应用于CDMA系统多用户检测,仿真表明:这种方法对远近问题不敏感,具有良好的误码率性能和抗多址干扰性能。Based on adaptive projective learning algorithm which is a simple and efficient means for constructing and training radial basis function neural networks, the centers of RBF and the weights of networks can be choosed one by one. In this paper, the RBF neural network based on adaptive projective algorithm has been used to solve the multiuser detection problem in CDMA. Simulation results are provided to show that this method is near-far resistant, and has comparable performance to the optimal detection approach.

关 键 词：自适应投影 学习算法 RBF神经网络 多用户检测方法 移动通信 码分多址 多址干扰 径向基函数 

分 类 号：TN929.533[电子电信—通信与信息系统] TP[电子电信—信息与通信工程]