Fusion Image Based Radar Signal Feature Extraction and Modulation Recognition

Abstract

The development of cognitive radio and radar electronic reconnaissance has put forward an important demand for improving the recognition ability of modulated signals in complex electromagnetic environment. In this paper, we propose a valid radar signal modulation recognition technology under low signal-to-noise ratio (SNR). The recognition technology can recognize 12 different modulation signals, including Costas, LFM, NLFM, BPSK, P1-P4, and T1-T4 codes. First, we propose the image fusion algorithm of non-multi-scale decomposition to fuse images of a single signal with different time-frequency (T-F) methods. Specifically, weights are designed by the principal component analysis, which could combine significative details of T-F images. Second, we adopt transfer learning-based convolutional neural networks and self-training-based stacked autoencoder, which extract the effective information on fusion image, furthermore guarantee the recognition performance. Moreover, multi-feature fusion algorithm is used to fuse features, which reduces redundant information on features and enhances computing efficiency. Finally, the classifier is performed by a classical algorithm called support vector machine. Simulation results show that the average recognition success rate is 95.5% at SNR of −6dB. It is testified that proposed recognition technology possesses good robustness and superiority in RSR with a wide range of SNR. INDEX TERMS Radar signal modulation recognition, time-frequency images, image fusion, multi-feature fusion.

Unknown Radar Waveform Recognition Based on Transferred Deep Learning

“…While, in the field of image classification, the bottom deep learning network can generalize common colors and features in natural images [29]. A variety of classical CNN are transferred to different classification tasks, and the training efficiency and classification accuracy of network models are greatly improved [30][31] [32]. Various typical CNN such as Alexnet [30] [31] and ResNet [32] have been enough trained maturely to be used in the classification of TFI of radar emitters.…”

“…A variety of classical CNN are transferred to different classification tasks, and the training efficiency and classification accuracy of network models are greatly improved [30][31] [32]. Various typical CNN such as Alexnet [30] [31] and ResNet [32] have been enough trained maturely to be used in the classification of TFI of radar emitters. In the literature [30], a mixture of pre-trained Encoder (SAE) and AlexNet structures have been used to extract TFI features for the situation where the radar signal sample is small.…”

“…Various typical CNN such as Alexnet [30] [31] and ResNet [32] have been enough trained maturely to be used in the classification of TFI of radar emitters. In the literature [30], a mixture of pre-trained Encoder (SAE) and AlexNet structures have been used to extract TFI features for the situation where the radar signal sample is small. It can identify 12 radar signals (Costas, LFM, NLFM, BPSK, P1-P4 and T1-T4) with an average recognition rate of 95.5 % when the SNR is greater than -6dB.…”

LPI Radar Waveform Recognition Based on Features from Multiple Images

“…In addition, there have been hybrid classifiers designed that combine a few different multiple classifiers. For instance, in [15,19,20], all of them combine two deep learning models. In [15], the classifier consists of CNN and ENN.…”

“…In [15], the classifier consists of CNN and ENN. In [20], the classifier consists of CNN and SAE.However, the feature of TFI will be overwhelmed by a large amount of noise due to the interference of noise. Researchers have dedicated their attention to image pre-processing, generally in two ways, in order to solve this problem.…”