提出了一种基于栈式自编码器与支持向量机的低截获概率（LPI）雷达信号识别方法。首先，通过Choi-Williams分布，将信号变换到时频域，获取信号的时频图像；其次，使用图像预处理方法对时频图像进行处理，得到便于自编码器处理的图像；再次，使用栈式自编码器从预处理后的时频图像中自动地提取出信号特征；最后，基于提取的信号特征使用支持向量机（SVM）对信号进行分类。本方法使用任意波形发生器（AWG）模拟产生了8类LPI雷达信号，采用栈式自编码器与支持向量机相结合的方法识别信号。仿真实验结果表明，该方法能够在低信噪比和小样本情形下有效识别LPI雷达信号。
A radar signal recognition method of low probability of intercept (LPI) based on the stacked autoencoder and the support vector machine is proposed. First, the signal is transformed to the time-frequency (T-F) domain to obtain the T-F images through the ChoiWilliams Distribution. Second, image processing methods are used to process the T-F images. Then, the stacked autoencoder is used to extract features from the preprocessed images. Finally, the support vector machine (SVM) is used to recognize the signal. The method uses the arbitrary waveform generator (AWG) to generate eight kinds of LPI radar signals and uses the stacked autoencoder combined with the SVM to recognize the signal. Simulation results show that the method can effectively classify the LPI radar signal in low SNR and small sample situations.
[1] Pace P E. Detecting and classifying low probability of intercept radar[M]. St Albans:Artech House, 2004:385-404.
[2] Raghu A S, Srikanth T, Kumar C S. Modulation classification of LPI radar using higher order statistics(HOS)[J]. Proc Spie, 2004, doi:10.1117/12.511179.
[3] Vanhoy G, Schucker T, Bose T. Classification of LPI radar signals using spectral correlation and support vector machines[J]. Analog integrated circuits and signal processing, 2017, doi:10.1007/s10470-017-0944-0.
[4] Gulum T O. Autonomous non-Linear classification of LPI radar signal modulations[D]. State of California:Monterey California Naval Postgraduate School, 2007.
[5] Zhang G, Dong Y, Liu P. Classification of low probability of interception communication signal modulations based on timefrequency analysis and artificial neural network[C]//Ningbo:International Conference on Electronics, Communications and Control, 2011:1936-1939.
[6] Zhang M, Liu L, Diao M. LPI radar waveform recognition based on Time-Frequency distribution[J]. Sensors, 2016, doi:10.3390/s16101682.
[7] Bai H, Zhao Y J, Hu D X. Radar signal recognition based on the local binary pattern feature of time-frequency image[J]. Journal of Astronautics, 2013, 34(1):139-146.
[8] Liu Y, Xiao P, Wu H, et al. LPI radar signal detection based on radial integration of Choi-Williams time-frequency image[J]. Journal of Systems Engineering and Electronics, 2015, 26(5):973-981.
[9] Liu G, Zhang G Y, Tian Z R. Radar signal recognition based on three dimensional entropy features of time-frequency image[J]. Fire Control & Command Control, 2016(12):173-177.
[10] Sun W, Shao S, Zhao R, et al. A sparse auto-encoder-based deep neural network approach for induction motor faults classification[J]. Measurement, 2016, 89:171-178.
[11] Cohen L. Time-frequency distributions:A review[J]. Proceedings of the IEEE, 1989, 77(7):941-981.
[12] Zilberman E R. Autonomous time-frequency cropping and feature-extraction algorithms for classification of LPI radar modulations[D], Monterey:Navel Postgraduate school, 2006.
[13] He K, Zhang X, Ren S, Sun J. Delving deep into rectifiers:Surpassing Human-Level performance on image net classification[C]//Santiago:Proceedings of the IEEE International Conference on Computer Vision, 2015:1026-1034.
