基于RBF神经网络的弱信号DOA估计方法

Weak signals' DOA estimation based on RBFNN

提出了一种利用智能学习方式解决强信号条件下弱信号的波达方向估计问题的新方法。通过理论推导发现协方差矩阵的特征值与来波信号功率之间存在一定的关系,将含功率信息的协方差矩阵的特征值加入到含方位信息的协方差矩阵上三角部分共同作为样本特征,构建基于RBF神经网络的弱信号DOA估计模型,解决了现有的智能学习方法难以对强信号条件下弱信号的来波方位估计问题,相比于利用解析计算的方法,本方法不需要已知强信号的来波方位并对其进行干扰抑制,计算量更小、DOA估计精度更高。仿真实验验证了新方法的有效性。

This paper proposed a novel intelligent technique for weak signals＇ DOA estimation in the presence of strong jam- ming or signal ,which transfered the problem of DOA estimation into a large mount of data intelligent learning and recognition problem. Firstly, obtained the relationship between eigenvalue of the correlation matrix and signal power according to the signal subspace theory. Extracted the upper triangular half of the correlation matrix and the eigenvalue of the correlation matrix of knowing direction signals to form training set together. Then constructed the weak signals＇ DOA estimation model based on RBFNN. Compared with the other methods of weak signals＇ DOA estimation using algebra calculation, the proposed approach needn＇ t to acquire the direction of the strong signals and attenuate them. In addition, the new method has less computing burden and higher estimation accuracy. The experiments demonstrate its effectiveness and feasibility.