摘要
针对通信辐射源个体识别面临噪声干扰与样本不足的问题,建立通信辐射源信号测量模型,开展小样本噪声条件下的信号识别研究。通过建立测量信号的时频谱统计量和时频能量谱统计量,分析累积时频能量谱的噪声抑制和特征收敛作用,提出基于时频域循环平稳特征图像的识别方法,构建用于深度学习的数据训练集和测试集,生成累积时频能量谱识别网络模型。实测数据和比对实验表明,网络模型在多周期累积后对信号时频特征图像的识别准确率达到90.1%,验证了所提出方法在噪声干扰和小样本数据条件下进行通信辐射源个体识别的适用性和稳定性。
In response to the problems of noise interference and insufficient samples in communication emitter individual identification,a communication emitter signal measurement model is established to conduct research on signal identification under small samples and noisy conditions.By establishing time-frequency spectrum statistics and time-frequency energy spectrum statistics of measurement signals,this paper analyzes the noise reduction and feature convergence effects of cumulative-cycle time-frequency energy spectrum,proposes a recognition method based on time-frequency domain cyclostationary feature images,constructs a data training set and test set for deep learning,and generates cumulative-cycle time-frequency energy spectrum recognition network models.The measured data and comparative experiments show that the network model achieves a recognition accuracy of 90.1%for signal time-frequency feature images after multiple cycles of accumulation,which verifies the applicability and stability of the proposed method for communication emitter individual identification under noise interference and small sample data conditions.
作者
黄宇
张鑫
于文龙
秦泗帅
林杭勇
HUANG Yu;ZHANG Xin;YU Wenlong;QIN Sishuai;LIN Hangyong(Unit 91715 of PLA,Guangzhou 510450,China;Naval Aviation University,Yantai 264001,China;Unit 91213 of PLA,Yantai 264000,China)
出处
《舰船电子对抗》
2025年第6期74-81,共8页
Shipboard Electronic Countermeasure
关键词
辐射源个体识别
小样本数据
累积时频能量谱
噪声抑制
emitter individual identification
small sample data
cumulative-cycle time-frequency energy spectrum
noise suppression
