基于FPGA的射频指纹特征提取系统设计  

System Design for RF Fingerprint Extraction Based on FPGA

作　　者：朱必新 王天雄 陈世霖 张炳坤 肖子谦 ZHU Bixin;WANG Tianxiong;CHEN Shilin;ZHANG Bingkun;XIAO Ziqian(National Key Laboratory of Wireless Communications,University of Electronic Science and Technology of China,Chengdu 611731,China;The 54th Research Institute of CECT,Shijiazhuang 050081,China)

机构地区：[1]电子科技大学通信抗干扰全国重点实验室,四川成都611731 [2]中国电子科技集团公司第五十四研究所,河北石家庄050081

出　　处：《无线电工程》2025年第4期898-904,共7页Radio Engineering

基　　金：国家自然科学基金(62106242)。

摘　　要：通信发射机中的模拟器件带来的信号畸变会使所发射的信号表征出发射机个体特征,这种特征通常称为射频指纹。蕴含在接收信号中的特征可以作为一种发射设备个体识别的手段。目前的射频指纹提取与识别方案大多依赖于复杂的信号预处理和计算量庞大的特征提取算法。在需要快速反应与决策的侦察系统中,由于计算资源有限,当前的系统设计很难达到实时性要求。为解决射频指纹识别系统在实际应用中的要求,提出了一种在非同步情况下的射频指纹个体识别方案及面向FPGA的实现结构,借助线性分类器验证了硬件设备的有效识别,在信噪比为0 dB的条件下,特征组合识别率均在92%以上。Signal distortion caused by the analog device will make the transmitted signal characterize the individual characteristics of the transmitter,which is called the radio frequency fingerprint.RF fingerprinting technology has been developed to identify the specific transmitter.At present,the implementation of RF fingerprint extraction and identification mostly depends on complex signal preprocessing and characteristic extraction algorithms with a huge amount of calculation.But in a reconnaissance system that requires rapid response and decision-making,due to limited computational resources,it is hard to meet real-time requirements.To address the requirements of radio frequency fingerprint identification systems in practical applications,a RF fingerprinting individual identification scheme is proposed for real-time requirement in non-synchronous situation and implanted based on FPGA,and the effective identification of hardware devices is verified with the help of linear classifiers.Under the condition of a signal-to-noise ratio of 0 dB,the identification rate of the characteristic combination is above 92%.

关 键 词：射频指纹 个体识别 实时处理 FPGA 

分 类 号：TN391.4[电子电信—物理电子学]