To address the significant performance degradation of domain adversarial neural network (DANN)-based specific emitter identification (SEI) algorithms under variable modulation scenarios, a VAE-DANN algorithm is proposed by reconstructing subtle features of radar signals through deep variational auto-encoder (VAE) optimization. Firstly, a deep VAE module is incorporated after the feature extraction module of DANN to construct the VAE-DANN architecture. Secondly, the cosine annealing strategy is applied to dynamically optimize weighting factors of the total loss function during adversarial training, where network parameters are adjusted by minimizing the total loss function to obtain the optimal model. Finally, SEI implementation for target-domain radar signals is accomplished under variable modulation conditions using the optimized VAE-DANN model. Experimental results demonstrated that the proposed algorithm achieved an average recognition accuracy exceeding 94% at 10 dB signal-to-noise ratio in variable modulation scenarios.
YAOQ, CHAIH, GAOM Y. Radar specific emitter identification using carrier frequency feature[C]∥Proceedings of the 2019 IEEE International Conference on Signal, Information and Data Processing. Piscataway,USA: IEEE, 2019. DOI: 10.1109/ICSIDP47821.2019.9173508 .
[3]
LINGQ, YANW J, ZHANGY C, et al. Transfer learning method for specific emitter identification based on pseudo-labelling and meta-learning[J]. IET Radar, Sonar & Navigation, 2024,18(9):1460-1473.
[4]
FANGY Y, WEIS, ZHAOY, et al. Radar-specific emitter identification with only envelope power based on multidimensional complex noncentral chi-square classifier[J]. IEEE Sensors Journal, 2023,23(17):20223-20235.
[5]
WONGL J, HEADLEYW C, MICHAELSA J. Specific emitter identification using convolutional neural network-based IQ imbalance estimators[J]. IEEE Access, 2019,7:33544-33555.
[6]
GANINY, LEMPITSKYV. Unsupervised domain adaptation by backpropagation[C]∥Proceedings of the 32nd International Conference on International Conference on Machine Learning. New York, USA: JMLR.org, 2015:1180-1189.
[7]
郑远浩.复杂场景下雷达辐射源个体识别技术研究[D].郑州:信息工程大学,2024:1-65.
[8]
ZHANGX L, LIT Y, GONGP, et al. Variable-modulation specific emitter identification with domain adaptation[J]. IEEE Transactions on Information Forensics and Security, 2022,18:380-395.
[9]
HUANGZ W, WANGL P, WUW M. DFNet: towards real-time semantic segmentation with depthwise separable convolution[C]∥Proceedings of the 2023 4th International Conference on Intelligent Design. Piscataway,USA: IEEE, 2023:230-235.
[10]
WANGZ Y, JIS W. Smoothed dilated convolutions for improved dense prediction[J]. Data Mining and Knowledge Discovery, 2021,35(4):1470-1496.