摘要
现有射频指纹识别研究主要集中于射频指纹产生及提取的通信机理,忽略了实际应用中信号采集数据的清洗筛分、识别算法模型的效率等工程性问题。为此,分析了卫星通信信号的载波信息提取的原理与方法,并针对现有射频指纹方法的不足,围绕卫星信号识别领域,利用海量采集的射频指纹数据,深入研究基于自组织神经网络的射频指纹识别算法,提出了对应的算法模型,并与现有常见的无监督算法进行了比较。结果表明,所提方法可以取得更优的算法聚类效果和时间开销,可作为设计实现卫星频谱管理系统的基础。
Current research on radio frequency(RF)fingerprint recognition mainly focuses on the communication mechanism of RF fingerprint generation and extraction,but ignores engineering problems such as the efficiency of data cleaning and recognition models in practical applications.Aiming at these shortcomings,this paper analyzed the principles and methods for extracting carrier signal information of satellite communication signals.Focusing on the field of satellite signal recognition,this paper used the massively collected RF fingerprint data,deeply studied the RF fingerprinting algorithm based on self-organizing neural networks,and proposed the corresponding algorithm model.In comparison with unsupervised algorithms,our proposed algorithm can achieve a higher clustering accuracy and lower time cost,and could be used as the basis for the design and implementation of a satellite spectrum management system.
作者
胡苏
马上
林迪
吴薇薇
HU Su;MA Shang;LIN Di;WU Weiwei(National Key Lab on Communications University of Electronic Science and Technology of China,Chengdu 611731,China)
出处
《信息对抗技术》
2022年第1期55-61,共7页
Information Countermeasures Technology
基金
国家自然科学基金资助项目(61971092)
四川省杰出青年科技人才基金资助项目(2020JDJQ0023)。
关键词
射频指纹
卫星通信
信号识别
聚类算法
RF fingerprint
satellite communication
signal identification
clustering algorithm
