基于脉冲神经网络的雷达辐射源调制类型识别

Modulation Pattern Recognition of Radar Emitter Based on Spiking Neural Network

面对日益复杂的电磁环境和层出不穷的新体制雷达,基于人工方式提取雷达辐射源特征难以满足现代认知电子战的需求。为提升雷达辐射源识别的智能化水平,提出一种新的基于脉冲神经网络(Spiking Neuron Network,SNN)进行雷达辐射源调制类型识别的算法。首先利用时频分析的方法,将5种常见雷达时域信号转换为二维灰度图,使用高斯调谐曲线编码器将输入数据转化为脉冲发放时刻,然后传入由Tempotron组成的脉冲神经网络进行识别。仿真实验结果表明脉冲神经网络具有优良的检测精度,功耗较低,验证了该方法的有效性。

In the face of increasingly complex electromagnetic environment and endless emergence of new radar systems,it is difficult to meet the needs of modern cognitive electronic warfare by extracting the characteristics of radar radiation sources based on artificial methods.In order to improve the intelligent level of radar radiation source identification,a new algorithm based on Spiking Neuron Network(SNN)is proposed for radiation source identification.First,the method of time-frequency analysis is used to convert five common radar time-domain signals into two-dimensional grayscale images,and a Gaussian tuning curve encoder is used to convert the input data into spike firing moments,and then they are passed into a SNN composed of Tempotron to be recognized.The simulation experiment results show that the SNN has good detection accuracy and low power consumption,which verifies the effectiveness of the method.