With the increase of radar signal modulations and the emergence of new system radars, the receiver will intercept radar signals at the same time. In order to accurately estimate and suppress the signals, this paper proposes an accurate recognition system for radar emitter signals. The system can effectively separate multiple signals and accurately recognize Binary Phase Shift Keying (BPSK), Linear Frequency Modulation (LFM), Continuous Wave (CW), Costas, Frank code, and P1 to P4 codes. The separation technique based on fractional Fourier transform is proposed to decompose received signals into multiple components.Furthermore, a transferable GoogleNet is explored to achieve accurate recognition of the first component with better separation effect. Meanwhile, variational mode decomposition is developed to eliminate the noise of the second component; then, the fusion features are extracted to improve the recognition rate of the second component. Finally, the improved particle swarm optimization algorithm is proposed to find best support vector machine parameters. The simulation results show that the recognition rate of single signal and double signals can reach 96.23% and 72%, respectively, when signal-to-noise ratio is 0 dB.The system can also bring some inspiration to medical and mechanical signal recognition.Trans Emerging Tel Tech. 2019;30:e3612.wileyonlinelibrary.com/journal/ett

The development of cognitive radio and electronic warfare brings new challenges to radar electronic reconnaissance, the recognition of radar signal plays an extreme important role in radar electronic reconnaissance. In order to realize the reliable recognition of radar signal at the condition of low signalto-noise ratio (SNR), we propose a new radar signal recognition system based on non-negative matrix factorization network (NMFN) and ensemble learning, which can recognize radar signals including BPSK, LFM, NLFM, COSTAS, FRANK, P1, P2, P3 and P4. First, we explore feature extractor based on convolutional neural network (CNN), which applies transfer learning to solve the problem of small sample size. Second, we propose non-negative matrix factorization network to extract features, which can reduce the redundant information. Third, we develop feature fusion algorithm based on stacked autoencoder (SAE), which can acquire essential expression of features and reduce dimension of features. Finally, we propose improved artificial bee colony algorithm (IABC) as the strategy of ensemble learning, which can improve the recognition rate. The simulation results show that the recognition rates reach 94.23% at −4 dB, 99.82% at 6 dB. INDEX TERMS Radar signal recognition, non-negative matrix factorization network, transfer learning, feature fusion, improved artificial bee colony algorithm.

Radar signal recognition is an indispensable part of an electronic countermeasure system. In order to solve the problem that the current techniques have, which is a low recognition rate and a slow recognition speed for radar signals, a rapid accurate recognition system is proposed, especially for when multiple signals arrive at the receiver. The proposed system can recognize eight types of radar signals while separating signals: binary phase shift keying (BPSK), linear frequency modulation (LFM), Costas, Frank code, and P1–P4 codes. Regression variational mode decomposition (RVMD) is explored to separate the received signals, which saves time for parameter optimization of variational mode decomposition (VMD). Furthermore, signal separation and a noise removal technique based on VMD and the first component recognition technique based on a deep belief network (DBN) are proposed. In addition, in order to overcome the loss of the secondary component caused by signal separation, a fusion network is explored to increase the recognition rate of the secondary component in a short time. The simulation results show that the recognition system achieves an overall recognition rate of 99.5% and 94% at a signal-to-noise ratio (SNR) of 0 dB when receiving single signals and double signals, while spending 0.8 s and 2.23 s, respectively. The proposed system can also be used to recognize medical and mechanical signals.

No abstract

No abstract