摘要
传统的基于密码机制和安全协议的无线网络安全存在隐患,新的基于物理层的射频指纹(radio frequency fingerprinting,RFF)方法利用发射机射频信号的细微差异来区分不同个体,具有难以克隆、伪造的优点,有着广阔的应用前景.本文首先讨论了理想RFF应具备的四种基本特性,即唯一性、时不变性、独立性和稳健性,分析了在四种基本特性方面的研究现状.然后按照信号预处理、特征提取和分类识别三个部分,对RFF识别相关技术进行了总结,重点分析了射频独特原生属性(RF-distinct native attribute,RF-DNA)、调制域和基于深度学习的RFF识别方法.最后,对RFF识别研究中涉及到的各种信号类型/调制方式及对应的应用场景进行了总结,展示了RFF识别的广阔应用前景,并对RFF识别的研究趋势进行了讨论.
The traditional wireless network security based on cryptographic mechanism and security protocol has potential safety hazard.The new physical layer-based RF fingerprinting method uses the fine-feature of transmitter RF signals to distinguish different individuals.It has the advantages of being difficult to clone and forge,and has broad application prospects.This paper first discusses the four basic characteristics of ideal RF fingerprinting,namely uniqueness,time invariance,independence and robustness,and analyzes the research status in four basic characteristics.Then according to the three parts of signal preprocessing,feature extraction and classification,the status of RF fingerprint identification is summarized.The RF-distinct native attribute(RF-DNA),modulation domain and RF fingerprinting method based on deep learning are analyzed.Finally,the various signal types/modulation methods and corresponding application scenarios involved in RF fingerprint identification research are summarized,and the broad application prospects of RF fingerprint recognition are presented.The research trend of RF fingerprint recognition is discussed.
作者
曾勇虎
陈翔
林云
郝晓军
许雄
汪连栋
ZENG Yonghu;CHEN Xiang;LIN Yun;HAO Xiaojun;XU Xiong;WANG Liandong(State Key Laboratory of Complex Electromagnetic Environment Effects on Electronics and Information System,Luoyang 471003,China;College of Information and Communication Engineering,Harbin Engineering University,Harbin 150001,China)
出处
《电波科学学报》
EI
CSCD
北大核心
2020年第3期305-315,共11页
Chinese Journal of Radio Science
基金
国家自然科学基金(61771154)。
