Intra-Pulse Modulation Classification of Radar Emitter Signals Based on a 1-D Selective Kernel Convolutional Neural Network

The intra-pulse modulation of radar emitter signals is a key feature for analyzing systems. Traditional methods which require tremendous amount prior knowledge are insufficient to accurately classify the modulations. Recently, deep learning-based methods, especially convolutional neural networks (CNN), have been used in classification signals. However, those two-dimensional CNN-based dimensional transformation original sampled stage data preprocessing, resource-consuming and poorly feasible. In order solve these problems, we proposed one-dimensional selective kernel network (1-D SKCNN) Compared with other previous described literature, preprocessing method merely includes zero-padding, fast Fourier (FFT) amplitude normalization, much faster easier achieve. experimental results indicate that has advantages speed higher accuracy