基于射频指纹的物联网设备身份识别方法  

Identification Method of Internet of Things Equipment Based on Radio Frequency Fingerprint

作　　者：刘念 任佳鑫 韩晓艺 刘铭 LIU Nian;REN Jia-xin;HAN Xiao-yi;LIU Ming(College of Computer and Information Technology,Beijing Jiaotong University,Beijing 100044)

机构地区：[1]北京交通大学计算机与信息技术学院,北京100044

出　　处：《现代计算机》2020年第19期3-8,21,共7页Modern Computer

基　　金：国家自然科学基金项目:基于设备指纹的无线物联网设备身份识别研究(No.61971029)。

摘　　要：现有物联网中通常采用单向鉴权方式,即侧重于中心节点鉴别物联网设备的身份真伪,存在一定的安全漏洞。为了解决该问题,提出一种适用于物联网场景的基于射频指纹技术的身份识别方法,使得物联网设备能有效识别其中心节点身份的真伪。具体的,通过利用射频链路双向通信中存在的互易规律,设计基于自动编码器的射频指纹转换器,将物联网设备所需完成的学习任务迁移到中心节点完成,以减小物联网设备侧的能量消耗。仿真证明在较低信噪比情况下该方法具有较好的身份识别性能。The existing Internet of Things usually adopts one-way authentication,which has certain security loopholes.To solve this problem,an identification method based on radio frequency fingerprint technology suitable for the Internet of Things scenario is proposed.This solution enables IoT devices to effectively identify the authenticity of their central nodes.Specifically,by utilizing the reciprocity law existing in the two-way communication of the radio frequency link,this paper designs a radio frequency fingerprint converter based on an AutoEncoder.The RF fingerprint converter can transfer the learning task of the IoT device to the central node to complete.This method can effectively reduce the energy consumption of IoT devices,and has better identification performance at lower signal-to-noise ratios.

关 键 词：射频指纹 身份识别 物联网安全 自动编码器 

分 类 号：TN915.05[电子电信—通信与信息系统] TP391.44[电子电信—信息与通信工程]