混沌背景下基于RBF神经网络的弱信号检测

RBF neural network-based weak signal detection in chaos

弱信号检测问题是目标检测中一个重要的研究内容。通常,采用贝叶斯(Bayes)方法来检测目标信号的存在。在本文中利用背景信号为混沌这一先验信息,采用了RBF神经网络对模拟产生的淹没在混沌背景中的暂态信号进行检测,并将该方法与采用BP神经网络时的检测性能进行了比较。仿真实验结果表明,基于RBF神经网络的检测性能优于BP神经网络。

The detection problem of the weak signal is an important aspect in the target detection. In general, the exist of the target is detected by using the Bayes' method. In this paper, the prior information that the background signal is in fact chaotic is exploited and a method for the detection of a transient signal buried in a chaotic background is presented using a RBF neural network. The results are compared to the method using a BP neural network, and they show that the method based on the RBF neural network has a better performance than the one based on the BP neural network.