摘要
以最小均方差(LMS)算法的解为基础,提出一种基于径向基函数(RBF)神经网络的波束形成方法。首先方法将一定量、等步长的DOA值通过LMS算法计算出的阵列权值向量作为训练集,而后基于训练集RBF神经网络进行训练,最终在DOA估计后可快速计算出阵列天线中各阵元的权值。上述方法利用RBF神经网络快速逼近非线性函数的特点,替代原有LMS自适应波束形成算法的迭代过程,以期减小计算量。Matlab仿真结果表明,上述方法能够有效减少阵列权值向量计算过程的运算量,提高波束形成的实时性。
A beamforming method based on radial basis function(RBF) neural network is proposed in this paper, which was realized with least mean square(LMS) algorithm. The method used a certain amount valid DOA values as the training set and the array weight vector calculated by the LMS algorithm to train the RBF neural network. The way to calculate the weight of each antenna array element after DOA estimation can be more quickly. It used the characteristic of RBF neural network that can approximate the nonlinear functions fast, and eliminates the iteration process of LMS adaptive beamforming algorithm. The method was simulated through MATLAB, and the result demonstrates the effectiveness of this method, reduces the computational complexity of the beamforming process and improves the real-time capability of beamforming.
作者
周书宇
黄宛宁
李崔春
ZHOU Shu-yu;HUANG Wan-ning;LI Cui-chun(Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China;School of Electronic Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China)
出处
《计算机仿真》
北大核心
2020年第10期159-163,共5页
Computer Simulation
基金
国家自然科学基金青年科学基金项目(61703389)。
关键词
波束形成
径向基函数
神经网络
最小均方算法
Beamforming
Radial basis function(RBF)
Neural networks
Least mean square(LMS)algorithm
