无人机由于体积小、成本低、灵活性高、机动性强等特点，已广泛应用于军用和民用领域。在军用领域，无人机主要执行危险性较高的任务，并带来了巨大的军事效益；在民用领域，无人机主要协助人们进行农业管理、电力巡检和地质勘查等工作。然而，在无人机应用范围不断扩大的同时，其面临的各类安全威胁也逐渐增多。其中，虚假欺骗攻击就是无人机面临的一种典型的网络安全威胁。
      目前，国内外针对无人机遭受的虚假欺骗攻击已经提出了一些解决方案。然而，通过已有研究分析发现，在利用深度学习技术检测无人机虚假欺骗攻击中的异常信号时，由于无人机异常信号与正常信号数据特征相似，加之采集的信号，会受到干扰影响，导致异常信号分类难度加大，从而影响到无人机异常信号检测准确度。为解决上述问题，本文以无人机 ADS-B 信号为研究对象，提出了基于卷积神经网络的无人机 ADS-B 异常信号检测方法。 其主要工作内容如下所述：
      (1) 为了解决因无人机正常信号和异常信号数据特征相似，导致异常信号分类难度加大，而影响无人机异常信号检测准确度的问题，本文提出了一种基于双捷径零偏置残差网络的无人机异常信号检测方法。该方法首先对无人机 ADS-B 信号进行去除随机噪声、数据归一化和傅立叶变换等数据预处理，得到频域信号；其次，构建一种新的双捷径残差网络，提取无人机频域信号特征；最后，通过零偏置全连接层与原有全连接层的融合，优化双捷径残差网络模型的最后一层网络结构，得到最终的双捷径零偏置残差网络，以便提高全连接层与 Softmax 层之间的关联度，并获得充分的无人机信号频域特征关联信息，从而提升无人机异常信号检测准确度。仿真结果表明，在使用数据质量较好的无人机 ADS-B 信号数据集时，该方法与现有异常信号检测方法相比，其检测精度有所提升。
      (2) 在真实场景下，收集的无人机信号不可避免地存在着大量随机噪声，导致无人机原始信号会受到干扰，从而影响无人机异常信号的检测准确度。因此， 本文提出了一种基于双模态感知的无人机异常信号优化检测方法。该方法首先对采集的无人机信号进行模态转化，生成波形图像；其次，利用双捷径零偏置残差网络，分别对无人机原始信号数据的波形图像模态和频域信号模态数据进行双模态的特征抽取；最后，利用提出的Bag-Stack 集成融合方式，将不同模态的数据分类结果融合，缓解随机噪声对无人机原始信号的干扰，以便在数据质量较差的条件下，也可以保证无人机异常信号检测准确度。仿真结果表明，在使用带有随机噪声的无人机原始 ADS-B 信号数据集时，该方法与现有方法相比，其检测精度有所提升。
      综上所述， 本文的开展和完成， 对于无人机异常信号检测具有一定的理论意义与实用价值。

     Unmanned Aerial Vehicles (UAVs) have been widely used in military and civilian fields due to their small size, low cost, high flexibility, and strong maneuverability. In the military fields, UAVs are mainly used for missions with high risk, bringing substantial military benefits. UAVs mainly assist people in daily tasks in the civilian feilds, such as farmland irrigation, urban management, and geological exploration. However, various security threats gradually increase by continuously expanding the scopes of UAVs' applications. Among all the security threats, false spoofing attack is a typical network threat faced by UAVs.

     At present, several solutions for solving the false spoofing attacks suffered by UAVs have been proposed at home and abroad. However, when detecting abnormal signals in false spoofing attacks of UAVs using deep learning, it is challenging to classify abnormal signals of UAVs due to similar data characteristics of abnormal signals and normal signals. Meanwhile, the collected signals of UAVs will be affected by interference, making it more challenging to classify abnormal signals for UAVs, thus affecting the detection accuracy of abnormal signals for UAVs. This thesis proposes abnormal signal detection methods for UAVs based on Convolutional Neural Networks to solve the problem. The main contents of this thesis are as follows:

     (1) In order to solve the problem that the detection accuracy of abnormal signal for UAVs is affected by the similar appearance and characteristics of normal signals data and abnormal signals data for UAVs, which makes it more challenging to classify abnormal signals. Therefore, this thesis proposes an abnormal signals detection method of UAVs based on Double Shortcuts Zero-Bias ResNet. First, the signals of UAVs are preprocessed by removing random noise, normalization, and Fourier transforms, then the frequency domain signals are obtained. Second, a new Double Shortcut Residual Network is constructed to extract characteristics of UAVs signals in the frequency domain. Finally, through the fusion of the Zero-Bias Fully Connected layer and the original Fully Connected layer, the network structure of the last layer of the Double Shortcuts ResNet model is optimized to obtain the final Double Shortcuts Zero-Bias ResNet, so as to improve the correlation degree of between the Fully Connected layer and Softmax layer. Simulation results indicate that when the data of ADS-B signals for UAVs set with good data quality is used, the proposed method can improve the detection accuracy compared with the existing abnormal signal detection methods.

     (2) In real scenes, there is inevitably a large amount of random noise in the collected UAV signals, which leads to interference of original signals of UAVs, thus affecting the detection accuracy of abnormal signals for UAVs. This thesis proposes an optimization detection method of abnormal signals of UAVs based on bimodal perception. First, the collected UAV signals are modally transformed to generate waveform images. Second, the Double Shortcuts Zero-Bias ResNet is used to extract bimodal features from the waveform image modal data and frequency domain signal modal data of original signals for UAVs. Then, the Bag-Stack ensemble fusion method is used to fuse the classification results of different modal data to alleviate the interference of random noise to the original signal of UAVs and improve the detection accuracy of abnormal signals on UAVs. Simulation results indicate that when the original ADS-B signal data set of UAVs with random noise is used, the proposed method can improve the detection accuracy compared with the existing abnormal signal detection methods.

     To sum up, the development and completion of this thesis have a particular theoretical significance and practical value for detecting abnormal signals of UAVs.

[1] 何道敬, 杜晓, 乔银荣, 朱耀康, 樊强, 罗旺. 无人机信息安全研究综述[J]. 计算机学报, 2019, 42(5): 1076–1094.

[2] N. Nagarani, P. Venkatakrishnan, and N. Balaji. Unmanned Aerial vehicle’s runway landing system with efficient target detection by using morphological fusion for military surveillance system[J]. Computer Communications, 2020, 151: 463–472.

[3] H. Shakhatreh, A. H. Sawalmeh, A. Al-Fuqaha, Z. Dou, E. Almaita, I. Khalil, N. S. Othman, A. Khreishah, and M. Guizani. Unmanned Aerial Vehicles (UAVs): A survey on civil applications and key research challenges[J]. IEEE Access, 2019, 7: 48572–48634.

[4] 王泽洋. 基于LSTM的无人机异常检测方法研究[D]. 黑龙江: 哈尔滨工业大学, 2018.

[5] O. Luňáček. Analysis of UAV usage in the physical security area[C]//2021 International Conference on Military Technologies. Brno, Czech Republic: IEEE, 2021: 1–5,

[6] R. Altawy and A. M. Youssef. Security, privacy, and safety aspects of civilian drones[J]. ACM Transactions on Cyber-Physical Systems, 2017, 1(2): 1–25.

[7] M. R. Manesh and N. Kaabouch. Cyber-attacks on unmanned aerial system networks: Detection, countermeasure, and future research directions[J]. Computers and Security, 2019, 85: 386–401.

[8] M. Fredrikson, S. Jha, and T. Ristenpart. Model inversion attacks that exploit confidence information and basic countermeasures[C]//the 22nd ACM SIGSAC Conference on Computer and Communications Security. Denver, USA: ACM, 2015: 1322–1333.

[9] M. Fang, X. Cao, , J. Jia, and N. Z. Gong. Local model poisoning attacks to byzantine-robust federated learning[C]//the 29th USENIX Security Symposium. Boston, MA, USA: USENIX, 2020: 1605–1622.

[10] S, Alfeld, X. Zhu, and P. Barford. Data poisoning attacks against autoregressive models[C]//AAAI Conference on Artificial Intelligence. Phoenix, Arizona: AAAI Press, 2016: 1452–1458.

[11] N. Akhtar and A. Mian. Threat of adversarial attacks on deep learning in computer vision: A survey[J]. IEEE Access, 2018, 6: 14410–14430.

[12] H. Sedjelmaci and S. M. Senouci. Cyber security methods for aerial vehicle networks: Taxonomy, challenges and solution[J]. The Journal of Supercomputing, 2018, 74(10): 4928–4944.

[13] N. M Rodday, R. D. O. Schmidt, and A. Pras. Exploring security vulnerabilities of unmanned aerial vehicles[C]//2016 IEEE/IFIP Network Operations and Management Symposium. Istanbul, Turkey: IEEE, 2016: 993–994.

[14] R. M. Fouda. Security vulnerabilities of cyberphysical unmanned aircraft systems[J]. IEEE Aerospace and Electronic Systems Magazine, 2018, 33(9): 4–17.

[15] M. Leonardi, L. D. Gregorio, and D. D. Fausto. Air traffic security: Aircraft classification using ADS-B message’s phase-pattern[J]. Aerospace, 2017, 4(51): 1–13.

[16] M. P. Arthur. Detecting signal spoofing and jamming attacks in UAV networks using a lightweight IDS[C]//2019 International Conference on Computer, Information and Telecommunication Systems. Beijing, China: IEEE, 2019: 1–5.

[17] H. Reyes and N. Kaabouch. Improving the reliability of unmanned aircraft system wireless communications through cognitive radio technology[J]. Communications and Network, 2013, 5(3): 225–230.

[18] Y. Zhang, J. Li, D. Zheng, P. Li, and Y. Tian. Privacy-preserving communication and power injection over vehicle networks and 5G smart grid slice[J]. Journal of Network and Computer Applications, 2018, 122: 50–60.

[19] J. A. Giraldo, D. I. Urbina, Á. A Cárdenas, J. Valente, M. A. Faisal, J. Ruths, N. O. Tippenhauer, H. Sandberg, and R. Candell. A survey of physics-based attack detection in cyber-physical systems[J]. ACM Computing Surveys, 2018, 51(4): 1–36.

[20] H. Sedjelmaci, S. M. Senouci, and N. Ansari. A hierarchical detection and response system to enhance security against lethal cyberattacks in UAV networks[J]. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2018, 48(9): 1594–1606.

[21] 王文益, 吴庆. 利用改进型AlexNet的ADS-B欺骗式干扰检测[J]. 信号处理, 2020, 36(5): 741–747.

[22] H. Yang, Q. Zhou, M. Yao, R. Lu, H. Li and X. Zhang. A practical and compatible cryptographic solution to ADS-B security[J]. IEEE Internet of Things Journal, 2019, 6(2): 3322–3334.

[23] 刘逸, 倪育德, 方姝, 崔瑞云, 王琳琳. 基于TDOA/TSOA的ADS-B虚假信号检测技术研究[J]. 现代导航, 2015, 6(2): 95–101.

[24] W. Wang, R. Wu, and J. Liang. ADS-B signal separation based on blind adaptive beamforming[J]. IEEE Transactions on Vehicular Technology, 2019, 68(7): 6547–6556.

[25] J. Naganawa, H. Tajima, H. Miyazaki, T. Koga, and C. Chomel. ADS-B antispoofing performance of monopulse technique with sector antennas[C]//2017 IEEE Conference on Antenna Measurements & Applications (CAMA). Tsukuba, Japan: IEEE, 2017: 87–90.

[26] 丁建立, 邹云开, 王静, 王怀超. 基于深度学习的ADS-B异常数据检测模型[J]. 航空学报, 2019, 40(12): 167–177.

[27] N. Ramdhan, M. Sliti, and N. Boudriga. Packet insertion attack detection in optical UAV networks[C]//2018 20th International Conference on Transparent Optical Networks (ICTON). Bucharest, Romania: IEEE, 2018: 1–5.

[28] E. Habler and A. Shabtai. Using LSTM encoder-decoder algorithm for detecting anomalous ADS-B messages[J]. Computers & Security, 2018, 78: 155–173.

[29] D. He, S. Chan, and M. Guizani. Drone-Assisted public safety networks: The security aspect[J]. IEEE Communications Magazine, 2017, 55(8): 218–223.

[30] Y. Javaid, W. Sun, and M. Alam. UAVSim: A simulation testbed for unmanned aerial vehicle network cyber security analysis[C]//2013 IEEE Globecom Workshops Wireless Networking and Control for Unmanned Autonomous Vehicles. Atlanta, GA, USA: IEEE, 2013: 1432–1436.

[31] M. T. Dabiri, M. J. Saber, and S. M. S. Sadough. On the performance of multiplexing FSO MIMO links in log-normal fading with pointing errors[J]. IEEE/OSA Journal of Optical Communications and Networking, 2017, 9(11): 974–983.

[32] H. Yang and Q. Geng. The design of flight control system for small UAV with static stability[C]//2011 Second International Conference on Mechanic Automation and Control Engineering. Inner Mongolia, China: IEEE, 2011: 799–803.

[33] D. Mendes, N. Ivaki, and H. Madeira. Effects of GPS spoofing on unmanned aerial vehicles[C]//2018 IEEE 23rd Pacific Rim International Symposium on Dependable Computing (PRDC). Taipei, Taiwan: IEEE, 2018: 155–160.

[34] D. He, S. Chan, and M. Guizani. Communication Security of Unmanned Aerial Vehicles[J]. IEEE Wireless Communications, 2017, 24(4): 134–139.

[35] T. P. Huster, C. J. Chiang, R. Chadha, and A. Swami. Towards the development of robust deep neural networks in adversarial settings[C]//2018 IEEE Military Communications Conference. Los Angeles, CA, USA: IEEE, 2018: 419–424.

[36] K. Wesson, M. Rothlisberger, and T. Humphreys. Practical cryptographic civil GPS signal authentication[J]. Navigation, 2012, 59(3): 177–193.

[37] J. Kerns, K. D. Wesson, and T. E. Humphreys. A blueprint for civil GPS navigation message authentication[C]//2014 IEEE/ION Position, Location and Navigation Symposium, Monterey, CA, USA: IEEE, 2014, 262–269.

[38] J. Magiera and R. Katulski. Detection and Mitigation of GPS Spoofing Based on Antenna Array Processing[J]. Journal of Applied Research and Technology, 2015, 13(1): 45–57.

[39] G. Panice, S. Luongo, G. Gigante, D. Pascarella, C. D. Benedetto, and A. Vozella. A SVM-based detection approach for GPS spoofing attacks to UAV[C]//2017 23rd International Conference on Automation and Computing. Huddersfield, UK: IEEE, 2017: 1–11.

[40] M. R. Manesh, J. Kenney, W. C. Hu, V. K. Devabhaktuni, and N. Kaabouch. Detection of GPS spoofing attacks on unmanned aerial systems [C]//2019 16th IEEE Annual Consumer Communications & Networking Conference. Las Vegas, NV, USA: IEEE, 2019: 1–6.

[41] K. Jansen, M. Schäfer, D. Moser, V. Lenders, C. Pöpper, and J. Schmitt. Crowd-GPS-Sec: Leveraging crowdsourcing to detect and localize GPS spoofing attacks[C]//2018 IEEE Symposium on Security and Privacy. San Francisco, CA, USA: IEEE, 2018: 1018–1031.

[42] Y. Zhi, Z. Fu, X. Sun, and J. Yu. Security and Privacy Issues of UAV: A Survey[J]. Mobile Networks and Applications, 2019, 25(1): 95–101.

[43] D. Davidson, H. Wu, R. Jellinek, T. Ristenpart, and V. Singh. Controlling UAVs with sensor input spoofing attacks[C]//the 10th USENIX Conference on Offensive Technologies. US: USWNIX Association, 2016: 221–231.

[44] 吴庆. 基于深度学习的ADS-B欺骗式干扰检测[D]. 天津: 中国民航大学, 2020.

[45] W. Wang, G. Chen, R. Wu, D. Lu, and L. Wang. A low-complexity spoofing detection and suppression approach for ADS-B[C]//2015 Integrated Communication, Navigation, and Surveillance Conference. Herdon, VA, USA: IEEE, 2015: K2-1–K2-8.

[46] 何梦林. 无人机的故障诊断与容错控制研究[D]. 贵阳: 贵州大学, 2015.

[47] 王泽洋. 基于LSTM的无人机异常检测方法研究[D]. 哈尔滨: 哈尔滨工业大学, 2018.

[48] Z. Birnbaum, A. Dolgikh, V. Skormin, E. O’Brien, D. Muller, and C. Stracquodaine. Unmanned Aerial Vehicle security using Recursive parameter estimation[C]// IEEE International Conference on Unmanned Aircraft Systems. Orlando, FL, USA: IEEE, 2014: 692–702.

[49] 吴飘. 基于深度学习的时序数据异常检测检测算法研究[D]. 大连: 大连理工大学, 2021.

[50] Taha and A. S. Hadi. Anomaly Detection Methods for Categorical Data: A Review[J]. ACM Computing Surveys, 2019, 52(2): 1–35.

[51] D. J. Hill and B. S. Minsker. Anomaly detection in streaming environmental sensor data: A data-driven modeling approach[J]. Environmental Modelling & Software, 2010, 25(9): 1014–1022.

[52] K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition[C]//2016 IEEE conference on computer vision and pattern recognition. Las Vegas, NV, USA: IEEE, 2016: 770–778.

[53] 马晗, 唐柔冰, 张义, 张巧灵. 语音识别研究综述[J]. 计算机系统应用, 2022, 31(1): 1–10.

[54] 奚雪峰, 周国栋. 面向自然语言处理的深度学习研究[J]. 自动化学报, 2016, 42(10): 1445–1465.

[55] 徐雄. 采用改进型AlexNet的辐射源目标个体识别方法[J]. 电讯技术, 2018, 58(6): 625–630.

[56] 韩萍, 孙丹丹. 特征选择与深度学习相结合的极化SAR图像分类[J]. 信号处理, 2019, 35(6): 972–978.

[57] 翁琳天然, 彭进霖, 何远, 钟都都, 彭建华, 峁旋宇. 基于深度残差网络的ADS-B信号辐射源个体识别[J]. 航空兵器, 2021, 28: 1–8.

[58] E. Wang, Y. Song, S. Xu, J. Guo, C. Hong, P. Qu, Z. Pang, and J. Zhang. ADS⁃B anomaly data detection model based on deep learning and difference of gaussian approach[J]. Transactions of Nanjing University of Aeronautics and Astronautics, 2020, 37(4): 550–561.

[59] L. J. Wong, W. C. Headley, S. Andrews, R. M. Gerdes, and A. J. Michaels. Clustering learned CNN features from raw I/Q data for emitter identification[C]//2018 IEEE Military Communications Conference (MILCOM). Los Angeles, CA, USA: IEEE, 2018: 26–33.

[60] 甄士博. 基于ADS-B技术的多无人机航迹规划算法研究[D]. 石家庄: 河北科技大学, 2018.

[61] 360独角兽安全团队. 无线电安全攻防大揭秘[M]. 北京：电子工业出版社, 2016: 78–88.

[62] K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV, USA: IEEE, 2016: 770–778.

[63] 刘冰, 李瑞麟, 封举富. 深度度量学习综述[J]. 智能系统学报, 2019, 14(06): 1064–1072.

[64] 王华军, 修乃华. 支持向量机损失函数分析[J]. 数学进展, 2021, 50(6): 801–828.

[65] W. Liu, Y. Wen, Z. Yu, M. Li, B. Raj, and L. Song. SphereFace: deep hypersphere embedding for face recognition[C]//The IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, USA: IEEE Press, 2017: 212–220.

[66] 史加荣, 王丹, 尚凡华, 张鹤于. 随机梯度下降算法研究进展[J]. 自动化学报, 2021, 47(9): 2103–2119.

[67] 李飞, 高晓光, 万开方. 基于权值动量的RBM加速学习算法研究[J]. 自动化学报, 2017, 43(7): 1142−1159.

[68] Z. Yu. Methodologies for cross-domain data fusion: an overview[J]. IEEE Transactions on Big Data, 2015, 1(1): 16–34.

[69] 余辉, 梁镇涛, 鄢宇晨. 多来源多模态数据融合与集成研究进展[J]. 情报理论与实践, 2020, 43(11): 169–178.

[70] D. Ramachandram and G. W. Taylor. Deep multimodal learning: a survey on recent advances and trends[J]．IEEE Signal Processing Magazine, 2017, 34(6): 96–108.

[71] 何俊, 张彩庆, 李小珍, 张德海. 面向深度学习的多模态融合技术研究综述[J]. 计算机工程, 2020, 46(5): 1–11.

[72] S. E. Kahou, C. PAL, X. Bouthillier, P. Froumenty, Ç. Gülçehre, R Memisevic, P. Vincent, A. Courville, and Y. Bengio. Combining modality specific deep neural networks for emotion recognition in video[C]//the 15th ACM on International Conference on Multimodal Interaction. New York, USA: ACM Press, 2013：543–550．

[73] 郭玥秀, 杨伟, 刘琦, 王玉. 残差网络研究综述[J]. 计算机应用研究, 2020, 37(5): 1293–1297.

[74] Y. Liu, J. Wang, J. Li, H. Song, T. Yu, S. Niu, and Z. Ming. Zero-bias deep learning for accurate identification of internet of things (IoT) devices[J]. IEEE Internet of Things Journal, 2021, 8(4): 2627–2634.

[75] 黄天镜. 基于多维特征的ADS-B数据异常检测方法研究[D]. 天津: 中国民航大学, 2020.

[76] 黄孝喜, 李晗雨, 王荣波, 王小华, 谌志群. 基于卷积神经网络与SVM分类器的隐喻识别[J]. 现代图书情报技术, 2018, 22(10): 77–83.

[77] 万磊, 佟鑫, 盛明伟, 秦洪德, 唐松奇. Softmax分类器深度学习图像分类方法应用综述[J]. 导航与控制, 2019, 18(6): 1–9.