一种基于新型RBF网络的侦察测向系统  

A Direction-Finding System Based on Novel RBF Networks

作　　者：宫新保[1] 臧小刚[1] 周希朗[1] 

机构地区：[1]上海交通大学电子工程系,上海200030

出　　处：《系统仿真学报》2004年第6期1175-1177,共3页Journal of System Simulation

基　　金：国防科技预研重点课题(33.6.2.7)

摘　　要：针对雷达侦察系统的到达角(AOA)测量问题,提出了一种新型的基于两级径向基函数(RBF)网络的侦察测向系统。根据RBF网络的特点,本文采用了一种基于免疫算法的混合方法确定RBF网络测向系统的参数。该方法利用免疫算法的全局搜索能力优化RBF网络隐层的结构和参数,并采用最小二乘算法计算RBF网络线性输出层的权值。计算机仿真表明,基于这种RBF网络的测向系统达到了很高的精度。A novel direction finding system based on two-layer radial basis function (RBF) network is proposed to solve the problem of AOA estimation problem in radar reconnaissance system. A hybrid method based on immune algorithm is used to determine the parameters of this RBF direction finding system. In the method structures and parameters of RBF hidden layer is optimized by immune algorithm with global searching ability, and weights of RBF linear output layer are computed with least square method. Computer simulations demonstrate that this RBF direction-finding system has reached very high precision.

关 键 词：测向 径向基函数网络 免疫算法 最小二乘法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TN95[自动化与计算机技术—控制科学与工程]