神经网络在无线电近炸引信信号处理中的应用

SUPPRESS BACKGROUND CLUTTER AND DETECT TARGET SIGNAL REAL TIME VIA NEURAL NETWORK IN PROXIMITY

本文给出了一种用神经网络实现的背景杂波实时抑制器。该抑制器根据对空引信的工作特点——目标尺寸小、信号持续时间短、背景回波功率高、信噪比低 ,利用神经网络 ,通过有效抑制背景杂波 ,提高信噪比 ,从而实现在强背景杂波中提取弱目标信号。本文采用具有分布结构的 BP网络实现神经网络。既便于硬件实现 ,又可以提高收敛速度 ,减少学习时间 。

A sea clutter suppresser for higher gains in signal-noise ratio is proposed for use in radio proximity fuze. It can precisely predict background clutter and is realized through the neural network having special construction. The neural network can greatly reduce the learning time, improve the convergence speed, and facilitate hardware realization. The background clutter suppresser is capable of processing signals in real time and has a very high gain in signal-noise ratio.