基于器件级仿真的用户设备射频指纹识别技术研究  

Research on RF Fingerprinting Technology of User Equipment Based on Device-level Simulation

作　　者：查浩然 王翰红 姜航 ZHA Haoran;WANG Hanhong;JIANG Hang(School of Information and Communication Harbin Engineering University,Harbin 150001,China;Key Laboratory of Advanced Marine Communication and Information Technology Ministry of Industry and Information Technology,Harbin Engiiieermg University,Harbin 150001,China)

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001 [2]哈尔滨工程大学先进船舶通信与信息技术工业和信息化部重点实验室,黑龙江哈尔滨150001

出　　处：《移动通信》2023年第2期46-51,64,共7页Mobile Communications

基　　金：国家自然科学基金面上项目(61771154)。

摘　　要：UE认证对6G无线通信安全至关重要,RFFI是指通过信号处理手段,对采集的UE信号进行特征提取,从而实现UE个体识别的一项技术。首先通过搭建器件级仿真平台,探讨NR系统中的发射机损伤对射频指纹的影响,包括振荡器偏移、IQ调制器偏移和功率放大器非线性。然后,采用不同的深度神经网络模型对仿真平台生成的数据进行仿真实验,结果表明,在发射机参数满足3GPP协议要求的前提下,对10个随机分布的射频参数的设备进行识别,当信噪比15 dB时识别准确率可以达到98%。UE authentication is crucial to the security of 6G wireless communication.RFFI refers to a technology of extracting features from collected UE signals through signal processing methods,so as to realize UE individual identification.This paper first discusses the impact of transmitter impairments on RF fingerprints in NR systems by building a devicelevel simulation platform,including oscillator offset,IQ modulator offset,and power amplifier(PA)nonlinearity.Then,different deep neural network models are used to simulate the data generated by the simulation platform.The results show that,on the premise that the transmitter parameters meet the requirements of the 3GPP protocol,10 devices with randomly distributed radio frequency parameters are identified and the recognition accuracy can reach 98%when the signal-to-noise is 15 dB.

关 键 词：射频指纹 器件级仿真 信号处理 

分 类 号：TN91[电子电信—通信与信息系统]