TY - JOUR
T1 - Exploiting deep learning for secure transmission in an underlay cognitive radio network
AU - Zhang, Miao
AU - Cumanan, Kanapathippillai
AU - Thiyagalingam, Jeyarajan
AU - Tang, Yanqun
AU - Wang, Wei
AU - Ding, Zhiguo
AU - Dobre, Octavia A.
N1 - Funding Information:
Manuscript received July 7, 2020; revised October 16, 2020 and December 10, 2020; accepted December 29, 2020. Date of publication January 8, 2021; date IRELESS communications have become an indispens-WofcurrentversionFebruary12,2021.TheworkofMiaoZhangwassupportedable part of daily life of people as they play a crucial Grant2020020070.TheworkofYanqunTangwassupportedinpartbytheby theResearch StartUpFundingofChongqingJiaotongUniversityunder role in our day-to-day activities and the means of interactions in Guangdong Natural Science Foundation under Grant 2019A1515011622 and in the current networked society. However, information security is partbytheNationalNaturalScienceFoundationunderGrant62071499.The one of the major challenges in wireless networks due to the open JiangsuProvinceunderGrantKTHY-039,inpartbytheScienceandTechnologyworkofWeiWangwassupportedinpartbytheSixCategoriesTalentPeakof nature of wireless signal transmission which is more vulnerable Program of Nantong under Grant MS22019019 and in part by the Verification for interception and eavesdropping. The conventional security Platform of Multi-tier Coverage Communication Network for oceans under methods employed at upper layers in the current communication Chen.(Correspondingauthor:YanqunTang.)GrantLZC0020.ThereviewofthisarticlewascoordinatedbyDr.Jung-Chieh systems completely rely on cryptographic techniques [1], [2]. Miao Zhang is with the School of Information Science and Engineer-Despite the fact that existing conventional security techniques, ing, Chongqing Jiaotong University, Chongqing 400074, China (e-mail: developed based on some high complex intractable mathemat-Kanapathippillai Cumanan is with the [email protected]). ical problems, are difficult to break or intercept, the broadcast Engineering, University of York, York, YO10 5DD, nature of wireless transmissions introduces different challenges [email protected]). in terms of secret key exchange and distributions [3], [4]. As a RutherfordAppletonLaboratory,ScienceandTechnologyFacilitiesCouncil,JeyarajanThiyagalingamiswiththeScientificComputingDepartmentof result, information theoretic based physical layer security has Harwell Campus, Ditcot OX11 0QX, U.K. (e-mail: [email protected]). been proposed to complement the conventional cryptographic YanqunTangiswiththeSchoolofInformationScienceandEngineering, methods and to provide additional security measures in wireless SchoolofElectronicsandCommunicationEngineering,SunYat-SenUniversity,ChongqingJiaotongUniversity,Chongqing400074,China,andalsowiththe transmissions. Furthermore, this approach exploits the dynamics Shenzhen 510 006, China (e-mail: [email protected]). of physical layer characteristics of wireless channels to es-WeiWangiswiththeSchoolofInformationScienceandTechnology,Nantong tablish secure transmission [1]. A reasonable secrecy rate can AdvancedCommunicationTechnologies,Nantong226001,China,andalsowithUniversity,Nantong226000,China,withtheNantongResearchInstitutefor be realized through physical layer security technique provided the Research Center of Networks and Communications, Peng Cheng Laboratory, that the signal-to-interference plus noise ratio (SINR) of the Shenzhen518000,China(e-mail:[email protected]). channel of the legitimate user is better than that of the channel ing,The Universityof Manchester,ManchesterM139PL, U.K.(e-mail:ZhiguoDingis withtheSchoolofElectricaland ElectronicEngineer- of the eavesdropper [5]. This novel technique was first theo- [email protected]). retically proved by Shannon [5] and then secrecy capacities OctaviaA.DobreiswiththeDepartmentofElectricalandComputerEn- of wiretap and related channels were developed by Wyner [6] [email protected]).gineering, Memorial University, St. John’s, NL A1B 3X5, Canada (e-mail: and Csiszar [7]. In contrast to the conventional cryptographic Digital Object Identifier 10.1109/TVT.2021.3050104 methods, physical layer security schemes are more suitable for 0018-9545 © 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Publisher Copyright:
© 1967-2012 IEEE.
PY - 2021/1
Y1 - 2021/1
N2 - This paper investigates a machine learning-based power allocation design for secure transmission in a cognitive radio (CR) network. In particular, a neural network (NN)-based approach is proposed to maximize the secrecy rate of the secondary receiver under the constraints of total transmit power of secondary transmitter, and the interference leakage to the primary receiver, within which three different regularization schemes are developed. The key advantage of the proposed algorithm over conventional approaches is the capability to solve the power allocation problem with both perfect and imperfect channel state information. In a conventional setting, two completely different optimization frameworks have to be designed, namely the robust and non-robust designs. Furthermore, conventional algorithms are often based on iterative techniques, and hence, they require a considerable number of iterations, rendering them less suitable in future wireless networks where there are very stringent delay constraints. To meet the unprecedented requirements of future ultra-reliable low-latency networks, we propose an NN-based approach that can determine the power allocation in a CR network with significantly reduced computational time and complexity. As this trained NN only requires a small number of linear operations to yield the required power allocations, the approach can also be extended to different delay sensitive applications and services in future wireless networks. When evaluate the proposed method versus conventional approaches, using a suitable test set, the proposed approach can achieve more than 94% of the secrecy rate performance with less than 1% computation time and more than 93% satisfaction of interference leakage constraints. These results are obtained with significant reduction in computational time, which we believe that it is suitable for future real-time wireless applications.
AB - This paper investigates a machine learning-based power allocation design for secure transmission in a cognitive radio (CR) network. In particular, a neural network (NN)-based approach is proposed to maximize the secrecy rate of the secondary receiver under the constraints of total transmit power of secondary transmitter, and the interference leakage to the primary receiver, within which three different regularization schemes are developed. The key advantage of the proposed algorithm over conventional approaches is the capability to solve the power allocation problem with both perfect and imperfect channel state information. In a conventional setting, two completely different optimization frameworks have to be designed, namely the robust and non-robust designs. Furthermore, conventional algorithms are often based on iterative techniques, and hence, they require a considerable number of iterations, rendering them less suitable in future wireless networks where there are very stringent delay constraints. To meet the unprecedented requirements of future ultra-reliable low-latency networks, we propose an NN-based approach that can determine the power allocation in a CR network with significantly reduced computational time and complexity. As this trained NN only requires a small number of linear operations to yield the required power allocations, the approach can also be extended to different delay sensitive applications and services in future wireless networks. When evaluate the proposed method versus conventional approaches, using a suitable test set, the proposed approach can achieve more than 94% of the secrecy rate performance with less than 1% computation time and more than 93% satisfaction of interference leakage constraints. These results are obtained with significant reduction in computational time, which we believe that it is suitable for future real-time wireless applications.
KW - Cognitive radio networks
KW - deep learning
KW - neural network
KW - physical layer security
KW - resource allocation techniques
UR - http://www.scopus.com/inward/record.url?scp=85099568428&partnerID=8YFLogxK
U2 - 10.1109/TVT.2021.3050104
DO - 10.1109/TVT.2021.3050104
M3 - Article
AN - SCOPUS:85099568428
SN - 0018-9545
VL - 70
SP - 726
EP - 741
JO - IEEE Transactions on Vehicular Technology
JF - IEEE Transactions on Vehicular Technology
IS - 1
M1 - 9318524
ER -