随着无线通信技术的不断发展，电磁环境的愈加复杂，通信过程不可避免地接收到无意和恶意干扰。其中，恶意干扰是破坏信息传输的人为辐射信号。若不能准确判断信号受扰状态，一方面通信系统无法采取有效抗干扰策略，进而无法保证通信过程的安全性及质量；另一方面不能根据受扰状态对设备进行有效的频谱管控，威胁设备的用频安全。因此，需要对信号受扰状态识别方法进行研究，从而为后续干扰抑制及频谱管控提供先验知识，以确保通信过程的顺利进行及电磁设备的用频秩序安全。

针对受扰状态识别，本文首先研究了数字通信信号特征参数随信噪比的变化情况。在此基础上，针对不同调制类型信号选取相应特征参数，本文提出基于支持向量机的受扰状态识别方法，实现了已知信号具体调制类型或仅知信号调制类别两种情况的受扰状态识别。同时，考虑到智能通信中信号调制类型会根据信道质量进行改变，本文还提出了基于残差神经网络的受扰状态识别方法。

本文主要内容可归纳为以下三部分：

（1）第二章介绍了通信信号受扰识别系统。首先，介绍了通信受扰过程，并对干扰信号分析系统进行概述。接着，对五种压制性干扰（STI、MTI、LFSI、NBI、PI）进行介绍及仿真。最后，分析了4类（PSK、ASK、FSK、QAM）共12种典型的数字调制信号及其受扰模型，从而为后续章节的受扰状态识别奠定基础。

（2）第三章研究了基于特征参数的受扰状态识别方法。在充分利用合作通信中已知信号调制类型和载频的先验条件下，首先，对12种数字调制信号的多种特征参数稳定性进行分析，确定用于受扰状态识别的特征参数及稳定区间。随后，对每种数字调制信号进行干扰仿真，分析所选特征参数对受扰状态识别的有效性。最后，针对已知信号具体调制类型和只能确定信号调制类别两种情况，提出基于决策树与支持向量机的受扰状态识别算法，并通过实验仿真验证了所提算法在低干信比下的有效性。

（3）第四章研究了基于残差神经网络的信号受扰状态识别算法，以应对智能通信中信号调制类型会随信道质量主动发生变化的情况。首先，介绍了卷积神经网络的基本原理。然后，使用残差神经网络实现对调制类型可变信号的受扰状态识别。最后，仿真结果表明，在比特信噪比5dB以上、干信比-5dB以上条件下，受扰状态识别准确率达96.5%，验证了所提算法的受扰状态识别有效性。

 

With the continuous development of wireless communication technology, the electromagnetic environment is becoming increasingly complex, and the communication process inevitably receives unintentional and malicious interference. Among them, malicious interference is the deliberate radiation signal that destroys information transmission. If the disturbed state of the signal cannot be accurately determined, on the one hand, the communication system cannot adopt effective anti-interference strategies, thus unable to guarantee the security and quality of the communication process; on the other hand, effective spectrum management of the equipment cannot be carried out according to the disturbed state, thus threatening the frequency safety of the equipment. Therefore, it is necessary to study the method of disturbed state identification of the signal, so as to provide prior knowledge for subsequent interference suppression and spectrum management, and ensure the smooth communication process and the frequency order safety of the electromagnetic equipment.

 

Regarding the recognition of disturbed states, this paper first studies the variation of characteristic parameters of digital communication signal with signal-to-noise ratio. Based on this, the corresponding characteristic parameters are selected for different modulation types of signals, and a disturbed state identification method based on Support Vector Machine (SVM) is proposed in this paper, which realizes the disturbed state identification for known signal modulation types or only known signal modulation categories. At the same time, considering that the signal modulation type in intelligent communication will change according to the channel quality, this paper also proposes a disturbed state identification method based on residual neural network.

 

The main content of this thesis can be summarized into the following three parts:

 

(1) Chapter 2 introduces the communication signal disturbed state identification system. Firstly, the communication interference process is introduced, and the interference signal analysis system is outlined. Then, five types of suppressive interference (STI, MTI, LFSI, NBI, PI) are introduced and simulated. Finally, 12 typical digital modulation signals of 4 categories (PSK, ASK, FSK, QAM) and their disturbed models are analyzed, laying the foundation for the subsequent chapters of disturbed state identification.

 

(2) Chapter 3 studies the disturbed state identification method based on characteristic parameters. Under the prior condition of making full use of the known signal modulation type and carrier frequency in cooperative communication, firstly, the stability of various characteristic parameters of 12 digital modulation signals is analyzed to determine the characteristic parameters and stable intervals used for disturbed state identification. Then, interference simulation is conducted for each type of digital modulation signals, and the effectiveness of the selected characteristic parameters for disturbed state identification is analyzed. Finally, for the two cases of known signal modulation type and only known signal modulation category, a disturbed state identification algorithm based on decision tree and SVM is proposed, and the effectiveness of the proposed algorithm is verified by experimental simulation under low interference-to-signal ratio..

 

(3) Chapter 4 studies the signal disturbed state identification algorithm based on residual neural network to deal with the situation where signal modulation type in intelligent communication actively changes with channel quality. Firstly, the basic principle of convolutional neural network is introduced. Then, residual neural network is used to identify the disturbed state of the signal with variable modulation type. Finally, the simulation results show that the accuracy of disturbed state identification is 96.5% under the conditions of bit signal-to-noise ratio above 5dB and interference-to-signal ratio above -5dB, which verifies the effectiveness of the proposed algorithm.

[1]赵福才, 孙银清. 美军商业卫星通信资源应用现状分析[J]. 卫星与网络, 2012, (8): 64-69
[2]曾祥明. 基于模式识别的通信对抗技术研究[D]. 重庆: 重庆大学, 2005.
[3]黄晓飞, 徐池. 卫星通信干扰样式研究[J]. 航天电子对抗. 2011, 027(006): 17-19.
[4]路远,李玉波, 凌永顺,等. 卫星信息对抗研究[J]. 现代防御技术. 2007, 35(5): 86-91.
[5]郭金梅. 卫星导航抗欺骗干扰技术综述[J]. 第十一届中国卫星导航年会论文集——S11 抗干扰与反欺骗技术, 2020.
[6]杨小牛. 通信电子战：信息化战争的战场网络杀手[M]. 北京: 电子工业出版社, 2011.
[7]王亚. 基于特征提取的多干扰检测与识别方法研究[D].成都: 电子科技大学,2022.DOI:10.27005/d.cnki.gdzku.2022.003517.
[8]Urkowitz H. Energy detection of unknown deterministic signals[J]. Proceedings of the IEEE, 1967, 55(4): 523-531.
[9]K Kyungtae, X Yan, R Sampath. Energy Detection Based Spectrum Sensing for Cognitive Radio: An Experimental Study[C]// Proceedings 2010 IEEE Global Communications Conference. USA: IEEE, 2010: 1-5.
[10]Chenyf. Improved energy detector for random signals in Gaussian noise[J]. IEEE Transactions on Wireless Communications, 2010, 9(2): 558-563.
[11]C Politis, S Maleki, J M Duncan. SDR Implementation of a Testbed for Real-Time Interference Detection With Signal Cancellation[J] . IEEE Access 2018, pp. 20808-20821.
[12]S Enseriank, D Cochran. A cyclostationary feature detector[C]// IEEE Conference Record of the Twenty-Eighth Asilomar Conference on Signals, Systems and Computers. Pacific Grove: IEEE, 1994, 806-810.
[13]J Lunden, S A Kassam, V Koivunen. Robust nonparametric cyclic correlation-based spectrum sensing for cognitive radio[J]. IEEE Transactions on Signal Processing, 2010, 58(1): 38-52
[14]王烁, 陈泽艺. 一种自适应干扰能量检测算法[J]. 安徽大学学报: 自然科学版, 2016, 40(1): 59-63.
[15]刘建成, 郝学坤, 王赛宇等. 基于频域中值滤波的自适应卫星通信恶意干扰检测[J]. 无线电通信技术, 2020, 46(3): 339-344.
[16]李奇, 徐慨, 蔡城鑫等. 基于 Welch 谱估计和能量检测 (ED) 的干扰信号检测技术[J]. 信息通信, 2018 (12): 66-70.
[17]Vartiainen J, Lehtomaki J J, Saarnisaari H. Double-threshold based narrowband signal extraction[C]//2005 IEEE 61st vehicular technology conference. IEEE, 2005, 2: 1288-1292.
[18]Lundén J, Koivunen V, Huttunen A, et al. Spectrum sensing in cognitive radios based on multiple cyclic frequencies[C]//2007 2nd International Conference on Cognitive Radio Oriented Wireless Networks and Communications. IEEE, 2007: 37-43.
[19]张天林. 基于信号能量的干扰检测技术研究[D]. 重庆大学, 2018.
[20]刘帅, 谭学治, 张南. 循环平稳特性检测频域设计及信噪比估计研究[J]. 电子技术应用, 2010 (2): 114-117.
[21]Ma S, Zhao C, Wang Y. Fractional low order cyclostationary spectrum sensing based on eigenvalue matrix in alpha-stable distribution noise[C]//2010 First International Conference on Pervasive Computing, Signal Processing and Applications. IEEE, 2010: 500-503.
[22]崔光辉. 基于时频分析的信号检测与参数估计方法研究[D]. 哈尔滨工程大学, 2015.
[23]王青媛, 张更新, 胡婧, 等. 一种基于时频分析的信号检测算法[J]. 无线电通信技术, 2016, 42(1): 38-42.
[24]Wu J, Guo F. Time-Frequency Parameter Estimation Method of Frequency Hopping Signal Based on Morphology Method under Low SNR[C]//2021 IEEE 6th International Conference on Signal and Image Processing (ICSIP). IEEE, 2021: 734-738.
[25]Yan Z, Xu Y, Wang L, et al. Feature extraction by enhanced time–frequency analysis method based on Vold-Kalman filter[J]. Measurement, 2023, 207: 112383.
[26]郑逸飞.基于分数阶时频变换的卫星导航干扰检测方法[D].合肥工业大学,2022.DOI:10.27101/d.cnki.ghfgu.2022.000842.
[27]康荣艳. 基于眼图分析的干扰信号特征建模与检测方法研究[D]. 西安: 西安电子科技大学,2021.
[28]Liu Y, Li Y, Zhu Y, et al. A brief review on deep learning in application of communication signal processing[C]//2020 IEEE 5th International Conference on Signal and Image Processing (ICSIP). IEEE, 2020: 51-54.
[29]Liedtke F F . Computer simulation of an automatic classification procedure for digitally modulated communication signals with unknown parameters[J]. Signal Processing, 1984, 6(4):311-323.
[30]Azzouz E E , Nandi A K . Automatic identification of digital modulation types[J]. Signal Processing, 1995, 47(1):55-69.
[31]Chan Y, Gadbois L, Yansouni P. Identification of the modulation type of a signal[C]// IEEE International Conference on Acoustics, Speech, & Signal Processing. IEEE, 2003.
[32]高新诚. 基于谱特征提取的调制识别算法研究[D]. 西安: 西安电子科技大学,2020.
[33]Reichert J. Automatic classification of communication signals using higher order statistics[C]//Acoustics, Speech, and Signal Processing, IEEE International Conference on. IEEE Computer Society, 1992, 5: 221-224.
[34]孙汝峰, 刘顺兰. 基于高阶累积量和谱分析的数字调制信号识别[J]. 微型机与应用, 2016, 35(4): 76-79.
[35]Spooner C M. Classificationb of co-channel communication signals using cyclic cumulants[C]//Conference Record of the Twenty-Ninth Asilomar Conference on Signals, Systems and Computers. IEEE, 1995, 1: 531-536.
[36]李越. 通信干扰样式识别与参数估计算法研究[D]. 西安: 西安电子科技大学, 2019.
[37]李奇,徐慨,杨海亮.干扰信号检测技术研究[J].信息通信,2018,0(6):28-31.
[38]Liang H, Ho K C. Identification of digital modulation types using the wavelet transform[C]// Military Communications Conference Proceedings, 1999. Milcom. IEEE Xplore, 1999:427-431 vol.1.
[39]Jin J D, Kwak Y J, Lee K W, et al. Modulation type classification method using wavelet transform for adaptive modulator[C]// International Symposium on Intelligent Signal Processing and Communication Systems. IEEE, 2005:289-292.
[40]张凯. 基于小波变换的时频分析方法研究及模块实现[D]. 成都: 电子科技大学,2022.
[41]闫文康, 闫毅, 范亚楠, 等. 基于小波变换熵值及高阶累积量联合的卫星信号调制识别算法[J]. 空间科学学报, 2021, 41(6): 968-975.
[42]赵宇峰, 曹玉健, 纪勇, 等. 基于循环频率特征的单信道混合通信信号的调制识别[J]. 电子与信息学报, 2014, 36(5): 1202-1208.
[43]Qi P, Zhou X, Zheng S, et al. Automatic modulation classification based on deep residual networks with multimodal information[J]. IEEE Transactions on Cognitive Communications and Networking, 2020, 7(1): 21-33.
[44]陈蕙心. 通信信号调制识别技术研究[D]. 西安: 西安电子科技大学, 2017.
[45]Hassan K, Dayoub I, Hamouda W, et al. Automatic modulation recognition using wavelet transform and neural network[C]//2010 9th International Conference on Intelligent Transport Systems Telecommunications,(ITST). IEEE, 2010: 234-238.
[46]孙丽英,陈滨,刘昌锦.基于深度学习的导航信号干扰存在性检测方法[J].佳木斯大学学报(自然科学版),2022,40(06):43-46.
[47]葛红妨.卫星干扰信号的检测与识别[D].杭州电子科技大学,2022.DOI:10.27075/d.cnki.ghzdc.2022.000556.
[48]乌倩倩. 基于深度学习的无线通信干扰信号检测[D].北京邮电大学,2020.DOI:10.26969/d.cnki.gbydu.2020.001422.
[49]冯小平, 李鹏, 杨绍全. 通信对抗原理[M]. 西安: 西安电子科技大学出版社, 2009.
[50]阙隆树, 冯全源. 数字调制信号识别中的相位去卷叠门限研究[J]. 信号处理, 2010, 26(1): 56-59.
[51]李一兵, 葛娟, 林云. 基于熵特征和支持向量机的调制识别方法[J]. 系统工程与电子技术, 2012.
[52]陈蕙心. 通信信号调制识别技术研究[D]. 西安: 西安电子科技大学, 2017
[53]张贤达. 时间序列分析: 高阶统计量方法[M]. 北京: 清华大学出版社, 1996.
[54]武亦文. 基于决策树的分类算法研究[J]. 数字通信世界, 2017, 12: 268-269.
[55]赖春廷. 决策树分类算法研究[J].信息与电脑(理论版),2020,32(14):59-62.
[56]梁欣萌. 北斗民用信号干扰检测与识别技术研究[D]. 天津: 中国民航大学, 2020.
[57]邓乃扬, 田英杰. 支持向量机: 理论, 算法与拓展[M]. 科学出版社, 2009.
[58]Wang G S, Ren Q H, Su Y Z. The interference classification and recognition based on SF-SVM algorithm[C]. 2017 9th IEEE International Conference on Communication Software and Networks, Guangzhou, 2017: 835-841.
[59]Kong M, Jing L, Zhang Z, et al. Radio ground-to-air interference signals recognition based on support vector machine[C]. 2015 IEEE International Conference on Digital Signal Processing (DSP), Singapore, 2015: 987-990.
[60]林明铨. 基于深度学习的WiFi干扰源识别[D]. 西安:西安电子科技大学,2021.
[61]王元东. 基于ResNet模型的图像分类方法及应用研究[D]. 南昌: 华东交通大学,2019.
[62]K He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C].Proceedings of
the IEEE conference on computer vision and pattern recognition. 2016: 770-778.
[63]王天昊. 一种改进的 CNN-ResNet 深度学习神经网络及应用[D]. 长春: 吉林大学,2020.