基于深度学习的蓝牙射频指纹识别系统  

Bluetooth Radio Frequency Fingerprint Identification System Based on Deep Learning

作　　者：陈拓 杨洁[1] 翟宇辰 安晨珲 李宗岩 CHEN Tuo;YANG Jie;ZHAI Yuchen;AN Chenhui;LI Zongyan(Nanjing Institute of Technology,Nanjing 210000,China)

机构地区：[1]南京工程学院,南京210000

出　　处：《移动信息》2023年第10期231-234,共4页MOBILE INFORMATION

摘　　要：蓝牙射频指纹具有难以伪造的优点,基于射频指纹的身份识别能有效提高网络的安全性。文中设计了一种基于深度学习网络的蓝牙射频指纹识别系统。首先,利用HackrfOne软件无线电平台和GNURadio软件在蓝牙信号广播阶段采集多种蓝牙信标信号。其次,对蓝牙信号进行预处理,将预处理后的数据分为训练集与验证集。然后,使用MATLAB深度学习工具箱来设计长短期记忆网,利用训练数据集对各个网络进行训练,得到蓝牙射频指纹识别网络。最后,利用验证集对上述网络进行测试和分析。当迭代次数为300时,网络对3种蓝牙信标的射频指纹识别的准确率均达到80%以上。Bluetooth RFID fingerprint has the advantage that it is difficult to forge,and the identification based on RF fingerprint can effectively improve the security of the network.A Bluetooth RF fingerprint identification system based on deep learning network is designed in this paper.First,use the Hackrf One software radio platform and GNU Radio software to collect a variety of Bluetooth beacon signals in the Bluetooth signal broadcasting stage.Secondly,the Bluetooth signal is preprocessed,and the preprocessed data is divided into training dataset and validation set.Then,the MATLAB deep learning toolbox is used to design the long short-term memory network,and the training data set is used to train each network to obtain the Bluetooth radio frequency fingerprint recognition network.Finally,the above network is tested and analyzed by using the validation set.When the number of iterations is 300,the accuracy rate of the radio frequency fingerprint recognition of the three Bluetooth beacons by the network reaches more than 80%.

关 键 词：蓝牙 软件无线电 射频指纹识别 深度学习 LSTM 

分 类 号：TN915.08[电子电信—通信与信息系统]