基于遗传RBF网络的混沌背景微弱信号检测方法

弱信号检测问题是目标检测中的一个重要研究内容。利用背景信号为混沌信号这一先验知识，采用径向基函数神经网络（RBFNN）建立混沌背景的一步预测模型，RBF网络利用遗传算法训练，最后，设计了门限滤波器，得到感兴趣的信号。仿真结果表明，在非线性程度很高的情况下，通过与其它几种训练RBF网络算法的比较，利用遗传算法训练的RBF网络具有最高的检测精度，并且训练得到的网络的复杂程度最低。

Method of weak signal detection embedded in chaos based on genetic RBF network

The problem of detecting weak signal is an important part in the target detection.According to the prior knowledge that the background signal is chaotic signal,this paper uses the radial basis function neural network to build one-step predictive model.RBF network is trained by genetic algorithm.In the end,a threshold filter is designed to get the interested signal.The results show that compared with several other algorithms used to train RBF network,RBF network trained by genetic algorithm has highest detection accuracy and lowest complicated network under non-linear circumstances.