Jamming is a big threat to radar system survival and anti-jamming is a part of the solution. The classification of radar jamming signal is the first step toward to anti-jamming. Recently, as an important part of deep learning, convolutional neural network (CNN) based methods have shown their capability in discriminant feature extraction and accurate classification. In this study, in order to harness the powerfulness of deep learning, CNN based methods are proposed to classify radar jamming signal acting on pulse compression radar. Specifically, a 1D-CNN is designed for radar jamming signal classification under the condition of sufficient training samples. Furthermore, due to the fact that the collection of sufficient training samples is time-consuming and expensive, a CNN-based siamese network is proposed for radar jamming signal classification to deal with the issue of limited training samples. The experimental results with sufficient and limited training samples show that the CNN-based classification methods obtain good classification performance in terms of classification accuracy and show a huge potential for radar jamming signal classification. INDEX TERMS Radar jamming signal, convolutional neural network (CNN), sufficient and limited training samples, siamese network.

The accurate classification of radar jamming signal is a core step of anti-jamming. Recently, convolutional neural network (CNN) based methods have shown their powerfulness in signal processing. In this paper, a deep fusion method based on CNN is proposed to classify jamming signal acting on pulse compression radar. The proposed method consists of three subnetworks (i.e., 1D-CNN, 2D-CNN, and fusion network). 1D-CNN is used to extract deep features of original radar jamming signal. Meanwhile, in order to extract the time-frequency features, short time Fourier transform (STFT) is applied to jamming signal to obtain time-frequency spectrograms. Then, 2D-CNN is used to extract deep time-frequency features, which are useful for further features fusion processing. Fusion network is used to deeply fuse the extracted features of the aforementioned CNNs and softmax is used to finish the task of radar jamming signal classification. In addition, in order to alleviate the problem of overfitting and improve the generalization ability of proposed model, soft label smoothing is proposed. The experimental results reveal that the proposed method provides competitive results in term of classification accuracy. INDEX TERMS Radar jamming signal, classification, convolutional neural network (CNN), deep fusion.