针对传统的基于卷积神经网络的通信调制识别方法存在的缺陷，本文提出一种基于卷积神经网络与对抗训练的通信调制识别方法。该方法首先构建卷积神经网络作为深度学习模型，然后通过卷积核提取调制信号的特征参数，并通过对抗训练的方法提升网络的抗噪性能。最后采用SoftMax层输出识别概率，达到多调制识别的目的。实验结果表明，在缺少信道和噪声等先验信息的条件下，该方法的识别率得到了进一步提升，能有效识别16QAM，64QAM等11种调制类别，具有较好的工程应用价值。
Aiming at the defects of traditional modulation recognition methods based on convolutional neural network, this paper proposes a modulation recognition method based on convolutional neural network and adversarial training. Firstly, the convolutional neural network is constructed as a deep learning model, and then the feature parameters of the modulation signal are extracted through the convolutional kernel, and the anti-noise performance of the network is improved by the adversarial training method. Finally, the recognition probability is output by SoftMax layer to achieve the purpose of multi-modulation recognition. The experimental results show that, in the absence of prior information such as channel and noise, the recognition rate of this method is further improved, and it can effectively identify 11 modulation categories such as 16QAM and 64QAM, which has good engineering application value.
2022,44(8): 122-126 收稿日期：2021-08-26
DOI：10.3404/j.issn.1672-7649.2022.08.025
分类号：U665.2
作者简介：王建雄(1987-),男,博士,工程师,研究方向为扩频信号处理
参考文献：
[1] DOBRE O A, BAR-NESS Y, SU W. Higher-order cyclic cumulants for high order modulation classification[C]//Military Communications Conference, Milcom '03. IEEE, 2003:112-117.
[2] WU H C, SAQUIB M, YUN Z. Novel automatic modulation classification using cumulant features for communications via multipath channels[J]. IEEE Transactions on Wireless Communications, 2008, 7(8):3098-3105
[3] 曾创展, 贾鑫, 朱卫纲. 通信信号调制方式识别方法综述[J]. 通信技术, 2015, 48(3):252-257
[4] GHANI N, LAMONTAGNE R. Neural networks applied to the classification of spectral features for automatic modulation recognition[C]//Military Communications Conference, MILCOM '93. Conference record. Communications on the Move, IEEE, 1993:111-115.
[5] DESIMIO M P, PRESCOTT G E. Adaptive generation of decision functions for classification of digitally modulated signals[C]//Aerospace and Electronics Conference, NAECON 1988. Proceedings of the IEEE 1988 National, IEEE, 1988:1010-1014.
[6] HUO X, DONOHO D. A simple and robust modulation classification method via counting[C]//IEEE International Conference on Acoustics, Speech and Signal Processing, IEEE, 1998:3289-3292.
[7] LOPATKA J, PEDZISZ M. Automatic modulation classification using statistical moments and a fuzzy classifier[C]//International Conference on Signal Processing Proceedings, IEEE, 2000:1500-1506.
[8] SWAMI A, SADLER B M. Hierarchical digital modulation classification using cumulants[J]. IEEE Translation and Communication, 2000, 48(3):416-429
[9] MALIATSOS K, VASSAKI S, CONSTANTINOU P. Interclass and intraclass modulation recognition using the wavelet transform[C]//IEEE, International Symposium on Personal, Indoor and Mobile Radio Communications, IEEE, 2007:1-5.
[10] FAISAL N K, ZHONG Kang-ping, ZHOU Xian. Joint OSNR monitoring and modulation format identification in digital coherent receivers using deep neural networks[J]. Optics Express, 2017, 25(15):17767
[11] O'SHEA T J, CORGAN J, CLANCY T C. Convolutional radio modulation recognition networks[C]//International Conference on Engineering Applications of Neural Networks, Springer International Publishing, 2016:213-226.
[12] KARRA K, KUZDEBA S, PETERSEN J. Modulation recognition using hierarchical deep neural networks[C]//IEEE International Symposium on Dynamic Spectrum Access Networks, IEEE, 2017:1-3.
[13] LI R, LI L, YANG S, et al. Robust automated vhf modulation recognition based on deep convolutional neural networks[J]. IEEE Communications Letters, 2018, PP(99):1-1
[14] 鲍丹. 非理想条件下的调制分类问题[D]. 西安:西安电子科技大学, 2007.
[15] GOODFELLOW I J, SHLENS J, SZEGEDY C. Explaining and harnessing adversarial examples[EB/OL]. http://arxiv.org/abs/1412.6572.
[16] HUAN C Y, POLYDOROS A. Likelihood methods for MPSK modulation classification[J]. IEEE Translation Communication, 1995, 43(2):1493-1504
[17] LONG C S, CHUGG K M, POLYDOROS A. Further results in likelihood classification of QAM signals[C]//Military Communications Conference, Milcom '94. Conference Record. IEEE, 1994:57-61.
[18] ALUISIO, FONTES, FUAD, et al. Automatic modulation classification using information theoretic similarity measures[J]. Communication and Computer, 2013(7):944-950
