基于多尺度融合注意力网络的无人机控制信号识别技术

doi:10.12305/j.issn.1001-506X.2026.10.37

系统工程与电子技术 ›› 2026, Vol. 48 ›› Issue (10): 3675-3682.doi: 10.12305/j.issn.1001-506X.2026.10.37

• 通信与网络 • 上一篇

基于多尺度融合注意力网络的无人机控制信号识别技术

郭嘉琦¹^,²(), 张伟²^,³, 李泽一¹^,², 李鹏飞¹^,², 张琬滢¹^,²

1. 中国电子科技集团公司第二十九研究所，四川成都 610036
2. 电磁空间安全全国重点实验室，四川成都 610036
3. 电子科技大学信息与通信工程学院，四川成都 611731

收稿日期:2025-09-05 出版日期:2026-10-25 发布日期:2026-09-30
通讯作者: 郭嘉琦 E-mail:2734896747@qq.com
作者简介:张　伟（1985—），男，研究员级高级工程师，博士，主要研究方向为电子对抗
李泽一（1994—），男，工程师，博士，主要研究方向为通信对抗
李鹏飞（1993—），男，高级工程师，博士，主要研究方向为电子对抗
张琬滢（2000—），女，硕士研究生，主要研究方向为电子对抗、信号与信息处理
基金资助:
国家自然科学基金（U23B2013）资助课题

Multi-scale fusion attention network for unmanned aerial vehicle control signal recognition technology

Jiaqi Guo¹^,²(), Wei Zhang²^,³, Zeyi Li¹^,², Pengfei Li¹^,², Wanying Zhang¹^,²

1. The 29th Research Institute of China Electronics Technology Group Corporation，Chengdu 610036，China
2. National Key Laboratory of Electromagnetic Space Security，Chengdu 610036，China
3. School of Information and Communication Engineering，University of Electronic Science and Technology of China，Chengdu 611731，China

Received:2025-09-05 Online:2026-10-25 Published:2026-09-30
Contact: Jiaqi Guo E-mail:2734896747@qq.com

摘要/Abstract

摘要：

随着恶意无人机威胁的日益严重，通过无线干扰技术破坏其控制信号成为重要的反制手段。然而，在复杂电磁环境中识别无人机的控制信号面临着众多噪声干扰的挑战。针对无人机控制信号在不同时间尺度上呈现的多层级特征结构特性，提出一种多尺度融合注意力网络（multi-scale fusion attention network, MSAN）。该方法通过并行感受野捕获和分层信号处理两个分支同时提取信号的微观、中观和宏观多尺度特征，利用卷积增强的多头自注意力机制建模长距离时序依赖关系，从而在低信噪比环境下有效抑制噪声干扰。实验结果表明，MSAN在−15 dB信噪比条件下对无人机控制信号的识别F1分数达到85.56%，相比当前最好的时间序列识别模型——时间二维变化模型提升了3.1个百分点，同时参数量减少了79.7%，为复杂电磁环境下的无人机控制信号高精度识别提供了轻量化解决方案。

关键词: 深度学习, 多尺度特征融合, 注意力机制, 无人机, 控制信号识别

Abstract:

With the increasing severity of malicious unmanned aerial vehicle (UAV) threats, disrupting their control signals through wireless jamming technology has become an important countermeasure. However, identifying control signals in complex electromagnetic environments faces the challenge of numerous noise interferences. This paper proposes a multi-scale fusion attention network （MSAN） based on the multi-level feature structure characteristics of UAV control signals across different time scales. The method simultaneously extracts micro, meso, and macro multi-scale features through two branches of parallel receptive field capture and hierarchical signal processing, and utilizes convolution-enhanced multi-head self-attention mechanism to model long-range temporal dependencies, thereby effectively suppressing noise interference in low signal-to-noise ratio environments. Experimental results show that MSAN achieves an F1 score of 85.56% for UAV control signal recognition under −15 dB signal-to-noise ratio conditions, improving by 3.1 percentage points compared to the current best time series recognition model——temporal 2D-variation modeling for general time series analysis, while reducing parameters by 79.7%, providing a lightweight solution for high-precision UAV control signal recognition in complex electromagnetic environments.

Key words: deep learning, multi-scale feature fusion, attention mechanism, unmanned aerial vehicle, control signal recognition

中图分类号:

TN 911

郭嘉琦, 张伟, 李泽一, 李鹏飞, 张琬滢. 基于多尺度融合注意力网络的无人机控制信号识别技术[J]. 系统工程与电子技术, 2026, 48(10): 3675-3682.

Jiaqi Guo, Wei Zhang, Zeyi Li, Pengfei Li, Wanying Zhang. Multi-scale fusion attention network for unmanned aerial vehicle control signal recognition technology[J]. Systems Engineering and Electronics, 2026, 48(10): 3675-3682.

图/表 5

图1

表1

图2

表2

表3

参考文献 31

1	Hassania S M, Hamood Q N O, Yanlong L, et al. Unmanned aerial vehicles （UAVs）: practical aspects, applications, open challenges, security issues, and future trends[J]. Intelligent Service Robotics, 2023, 16 (1): 21. doi: 10.1007/s11370-022-00452-4
2	Jalil H H, Yue C, Un K N, et al. A comprehensive survey on security, privacy issues and emerging defence technologies for UAVs[J]. Journal of Network and Computer Applications, 2023, 213, 103607. doi: 10.1016/j.jnca.2023.103607
3	Wang W X, Jiang B, Tan X D, et al. UAV communication interference analysis and anti-interference methods[C]// International Conference on Cyberspace Innovation of Advanced Technologies, 2021: 197.
4	苏志刚, 晏翔, 韩冰, 等. 低信噪比条件下无人机射频信号实时检测方法[J]. 信号处理, 2023, 39 (5): 919. doi: 10.16798/j.issn.1003-0530.2023.05.016
5	Wang X B, Ying T, Tian W, et al. Spectrum representation based on STFT[C]// 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics, 2020: 435.
6	蒋平. 基于信号特征的无人机航测技术研究及实现[D]. 桂林: 桂林电子科技大学, 2021.
7	王安平, 吕振彬, 陈祥轩, 等. 基于图像分割和2次分类的无人机信号识别方法[J]. 上海大学学报（自然科学版）, 2024, 30 (3): 435.
8	Ezuma M, Erden F, Anjinappa K C, et al. Detection and classification of UAVs using RF fingerprints in the presence of Wi-Fi and Bluetooth interference[J]. IEEE Open Journal of the Communications Society, 2020, 1, 160. doi: 10.1109/ojcoms.2019.2955889
9	Al-Sad F M, Al-Ali A, Mohamed A, et al. RF-based drone detection and identification using deep learning approaches: an initiative towards a large open source drone database[J]. Future Generation Computer Systems, 2019, 100, 86. doi: 10.1016/j.future.2019.05.007
10	Janeth A G, Ali A, Jae-Min L, et al. CNN-32DC: an improved radar-based drone recognition system based on convolutional neural network[J]. ICT Express, 2022, 8 (4): 606. doi: 10.1016/j.icte.2022.04.012
11	Fu R, Al-Absi M A, Kim K H, et al. Deep learning-based drone classification using radar cross section signatures at mmWave frequencies[J]. IEEE Access, 2021, 9, 161431. doi: 10.1109/ACCESS.2021.3115805
12	Flores A, Flores G. Transition control of a tail-sitter UAV using recurrent neural networks[C]// International Conference on Unmanned Aircraft Systems, 2020: 303.
13	孟宪鹏, 刘利民, 董健, 等. 基于双向循环神经网络的雷达波束发射方向识别[J]. 系统工程与电子技术, 2024, 46 (3): 898.
14	Zhou S H, He Z, Chen X, et al. An anomaly detection method for UAV based on wavelet decomposition and stacked denoising autoencoder[J]. Aerospace, 2024, 115 (5): 393. doi: 10.3390/aerospace11050393
15	Allahham M S, Khattab T, Mohamed A. Deep learning for RF-based drone detection and identification: a multi-channel 1-D convolutional neural networks approach[C]// IEEE International Conference on Informatics, IoT, and Enabling Technologies, 2020: 112.
16	李彬, 徐怡松, 罗杰. 采用残差神经网络的无人机航迹目标识别算法[J]. 西安交通大学学报, 2021, 55 (12): 146.
17	杨小伟, 文清丰, 杨雪, 等. 基于卷积神经网络的无人机航迹目标识别[J]. 无线电工程, 2022, 52 (3): 456.
18	Guo H L, Shi L L, Liu J L. An improved multi-scale feature extraction network for medical image segmentation[J]. Quantitative Imaging in Medicine and Surgery, 2024, 14 (12): 8331. doi: 10.21037/qims-24-1022
19	Zhang C L, Cao K, Lu L M, et al. A multi-scale feature extraction fusion model for human activity recognition[J]. Scientific Reports, 2022, 12 (1): 20620. doi: 10.1038/s41598-022-24887-y
20	Liao N D, Guan J Y. Multi-scale convolutional feature fusion network based on attention mechanism for IoT traffic classification[J]. International Journal of Computational Intelligence Systems, 2024, 17 (1): 36. doi: 10.1007/s44196-024-00421-y
21	Sinha A, Dolz J. Multi-scale self-guided attention for medical image segmentation[J]. IEEE Journal of Biomedical and Health Informatics, 2020, 25 (1): 121.
22	Wang M T, Fang S L, Fan Y C, et al. An ultra lightweight neural network for automatic modulation classification in drone communications[J]. Scientific Reports, 2024, 14 (1): 21540. doi: 10.1038/s41598-024-72867-1
23	Xiao Y, Zhang X J. Micro-UAV detection and identification based on radio frequency signature[C]// 6th International Conference on Systems and Informatics, 2019: 1056.
24	Yang S B, Luo Y, Miao W, et al. RF signal-based UAV detection and mode classification: a joint feature engineering generator and multi-channel deep neural network approach[J]. Entropy, 2021, 23 (12): 1678. doi: 10.3390/e23121678
25	Bithas S P, Michailidis T E, Nomikos N, et al. A survey on machine-learning techniques for UAV-based communications[J]. Sensors, 2019, 19 (23): 5170. doi: 10.3390/s19235170
26	张剑, 田福庆, 梁伟刚, 等. 基于多尺度AlexNet网络的健康因子构建方法[J]. 系统工程与电子技术, 2020, 42 (1): 245.
27	Wu D, Liu H N, Xu J W, et al. Multi-scale feature extraction for energy-efficient object detection in remote sensing images[J]. IET Computer Vision, 2024, 18 (8): 1223. doi: 10.1049/cvi2.12317
28	Sokolova M, Lapalme G. A systematic analysis of performance measures for classification tasks[J]. Information Processing & Management, 2009, 45 (4): 427. doi: 10.1016/j.ipm.2009.03.002
29	Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[C]// Advances in Neural Information Processing Systems, 2017.
30	Wu H X, Hu T G, Liu Y, et al. TimesNet: temporal 2D-variation modeling for general time series analysis[EB/OL]. [2025-07-05]. https: //arxiv.org/abs/2210.02186.
31	Zhou H Y, Zhang S H, Peng J Q, et al. Informer: beyond efficient transformer for long sequence time-series forecasting[C]// AAAI Conference on Artificial Intelligence, 2021.

信噪比/dB	精确率	召回率	F1分数
−15	88.90	82.47	85.56
−13	95.50	78.56	86.21
−11	95.53	81.08	87.71
−9	88.79	92.58	90.65
−7	89.79	97.72	93.59
−5	97.86	96.09	96.97
−3	97.81	98.72	98.26
−1	98.85	97.76	98.30
1	97.85	99.50	98.67
3	98.11	99.97	99.03
5	98.75	99.42	99.08

模型	识别率/%	参数量/MB
移除多尺度特征提取部分	74.21	1.234
移除特征融合部分	79.56	1.756
移除时序注意力编码部分	71.03	0.847
替换时序注意力编码部分	83.21	1.902
MSAN	85.56	1.902

[1]	马治勋, 汤宁, 李璇, 郝程鹏. 基于混合扩散模型的水声数据增扩方法[J]. 系统工程与电子技术, 2026, 48(9): 2898-2906.
[2]	魏帅, 谢荣, 杨洪飞, 许述文, 刘峥. 基于多尺度加权GCN的有源干扰识别方法[J]. 系统工程与电子技术, 2026, 48(9): 2958-2968.
[3]	王彩云, 韩嘉轩, 贾一帆, 王佳宁, 李晓飞. 多核保持分析与坐标关联注意的多源融合识别[J]. 系统工程与电子技术, 2026, 48(9): 2979-2986.
[4]	朱彦伟, 陈劲润, 赵晓锋, 龙洗, 乔琛远. 在轨目标探测任务-资源快速匹配方法[J]. 系统工程与电子技术, 2026, 48(9): 3029-3037.
[5]	邢娜, 唐旭莹, 王月海, 王婕, 宁可庆. 基于A^*引导的高斯伪谱法无人机协同轨迹规划[J]. 系统工程与电子技术, 2026, 48(9): 3210-3219.
[6]	黄敏, 李召瑞, 孙慧贤, 倪想. 基于动态加权均匀圆阵的无人机鲁棒测向方法[J]. 系统工程与电子技术, 2026, 48(8): 2520-2525.
[7]	陈任飞, 陈佳慧, 李忠文, 王思霖. 基于多相机信息协同的水面漂浮物跟踪方法研究[J]. 系统工程与电子技术, 2026, 48(8): 2545-2559.
[8]	班宏辉, 潘继飞, 王正, 崔瑞. 基于多无人机协同的航迹欺骗误差分析[J]. 系统工程与电子技术, 2026, 48(7): 2195-2205.
[9]	韩成杰, 彭聪, 施苏牧, 王宇, 纪江南. 基于多尺度交互融合网络的作动器故障诊断[J]. 系统工程与电子技术, 2026, 48(7): 2319-2332.
[10]	许贝佳, 刘学达, 禹明刚, 章乐贵. 基于多策略公共物品博弈的无人机集群弹药投放[J]. 系统工程与电子技术, 2026, 48(7): 2340-2351.
[11]	关巍, 胡彤博, 张显库, 罗春奇, 曲胜, 崔哲闻. 基于改进PPO算法的无人机/无人船协同导航方法[J]. 系统工程与电子技术, 2026, 48(7): 2457-2468.
[12]	黄琼丹, 刘露露, 韩洁婧, 王佳鹏, 康仕林. 基于差异特征反投影融合的图像超分辨重建[J]. 系统工程与电子技术, 2026, 48(6): 1809-1818.
[13]	袁胜智, 王少蕾, 李静, 肖任楷, 谭浩天, 赖建伟, 许江涛. 基于小波变换和改进残差神经网络的故障诊断方法[J]. 系统工程与电子技术, 2026, 48(6): 2031-2041.
[14]	刘瑞航, 刘海颖, 刘宇辰, 陈晨, 李铁香. 基于多特征正交约束的无人机跨视角地理定位方法[J]. 系统工程与电子技术, 2026, 48(6): 2072-2080.
[15]	张国庆, 姚桂鹏, 李纪强. 无人机/船协同降落的鲁棒自适应模糊控制[J]. 系统工程与电子技术, 2026, 48(6): 2081-2088.

基于多尺度融合注意力网络的无人机控制信号识别技术

Multi-scale fusion attention network for unmanned aerial vehicle control signal recognition technology

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 5

参考文献 31

相关文章 15

编辑推荐

Metrics

本文评价