摘要
提出了一种新的基于径向基(RBF)神经网络的相关干涉仪测向方法,实现了自组织学习选取中心、正交最小二乘法及基于遗传算法的进化优选算法等训练方法,经训练后的RBF神经网络可用于多源信号波达角(DOA)估计。仿真结果表明,在一定范围内,该方法对信道噪声不敏感,测向精度与传统相关干涉仪相当,且测向处理时间和测向设备的存储量大大降低。
In this paper,a new direction of arrival（DOA） estimation method of correlative interferometer based on RBF neural network is presented,and self-organized center selecting algorithm,orthogonal least squares method and evolutionary selecting algorithm based on genetic algorithm are realized.The computer simulation result shows that,for a certain extent,this new method is not sensitive to noise,the direction finding precision based on RBF neural network is the same as that traditional correlative interferometer,and the direction finding processing time and the storage space of direction finding are greatly reduced.
出处
《无线电工程》
2011年第1期15-17,50,共4页
Radio Engineering
关键词
相关干涉仪
测向
径向基函数神经网络
correlative interferometer
direction finding
radial basis function neural network
