Reliable resource allocation with RF fingerprinting authentication in secure IoT networks

作     者：Weiwei WU Su HU Di LIN Gang WU Weiwei WU;Su HU;Di LIN;Gang WU

作者机构：National Key Laboratory of Science and Technology on CommunicationsUniversity of Electronic Science and Technology of China 

出 版 物：《Science China(Information Sciences)》 (中国科学:信息科学(英文版))

年 卷 期：2022年第65卷第7期

页      面：51-66页

核心收录：

学科分类：1305[艺术学-设计学（可授艺术学、工学学位）] 13[艺术学] 0839[工学-网络空间安全] 081104[工学-模式识别与智能系统] 08[工学] 0804[工学-仪器科学与技术] 081101[工学-控制理论与控制工程] 0811[工学-控制科学与工程] 

基　　金：partially supported by Science and Technology Program of Sichuan Province (Grant No.2021YFG0330) Intelligent Terminal Key Laboratory of Si Chuan Province (Grant No. SCITLAB-0001) Fundamental Research Funds for the Central Universities (Grant No. ZYGX2019J076) National Natural Science Foundation of China (Grant No.61971092) Province Sichuan Foundation for Distinguished Young Scholars (Grant No. 2020JDJQ0023) 

主　　题：user authentication Internet of things convolutional neural network RF fingerprinting NP-hard optimization problem 

摘      要：The unprecedented growth of the Internet of Things(Io T) has led to a huge amount of wireless resource consumption in a network. Due to limited wireless resources, a network can only guarantee the quality of service(QoS) of authenticated users rather than that of all users. By acknowledging this limitation,we realise that user authentication would be a big issue in Io T networks. Although traditional authentication methods can enhance network security to a certain extent, their vulnerability to malicious attacks and the relevant complicated computations restrict Io T deployments. In this paper, a radio frequency(RF)fingerprinting based authentication scheme is proposed under the architecture of convolutional neural network(CNN). It can effectively prevent unauthenticated users from consuming valuable wireless resources and significantly improve QoS performance for legitimate users. By solving an NP-hard optimization problem with the objective of minimizing efficient energy density, we demonstrate an approximate optimal resource allocation scheme in consideration of an RF-fingerprinting based authentication process. The analytic results show that our proposed scheme can dramatically reduce the efficient energy density compared with traditional cryptography based authentication schemes.