Research On Physical Layer Security Technology Based On Collaboration

5G communication puts forward the requirements of high-speed,low-latency,extremely high reliability and high security,and proposes three application scenarios,i.e.,e MBB(high traffic mobile broadband),m MTC(large scale Internet),and u RLLC(low delay,high reliable connection).In addition,due to the flexible deployment of ultra-dense heterogeneous network and 100 billion level wireless sensor nodes with limited computational resources,it is difficult for the system to implement the complex key management,distribution and update processes among users,and increase the communication transmission delay.Under the classical password-based security mechanism,high security requirements at high speed and low latency are almost impossible to achieve.Relying solely on computational security-based cryptographic algorithms,it is difficult to meet the security requirements of diversified application scenarios in the future wireless network and high-quality user experience.Therefore,information security issues face very serious challenges.As a security complementary strategy to the classical cryptographic encryption mechanism,physical layer security has attracted extensive attention in recent years.Unlike password-based encryption technology,physical layer security does not assume that the eavesdropping node has limited computing power and cannot depend on the strength of the algorithm,which meets Shannon's “one-time pad” security condition.The physical layer security methods make full use of the random characteristics of wireless channels(e.g.,random fading,signal interference,and environmental noise)and advanced signal processing techniques(e.g.,artificial noise,cooperative jamming,and secure precoding).Meanwhile,physical layer security cannot require key distribution,which greatly saves system overhead,realize low latency,high security network access,and encrypted transmission.5G communication transmits from 3.5GHz to 6GHz,which is the best channel-sensitive characteristic frequency band for realizing physical layer security.The key technology of 5G communication using large-scale antennas also provides natural resources for physical layer security technology.On the basis of multiantenna secure transmission,the cooperation of multiple communication nodes or different service signals can further facilitate secure transmission.The power coordination of the transmitting nodes in cooperative secure transmission provides the possibilit the possibility for the implementation of the physical layer security.The cooperative secure transmission scenarios discussed in this dissertation mainly include cooperative jamming communication and friendly jammer serving other business subscribers of the system.Meanwhile,the adaptive secure transmission of confidential data to multiple legitimate users is an important security issue that needs to be solved urgently.For cooperative secure communication and secure multicast transmission scenarios,the research contents of this dissertation can be summarized as the following four parts.(1)Aiming at the assumption of channel estimation errors for legitimate channels in cooperative jamming communication scenarios,this dissertation investigates the effect of channel estimation errors on secrecy capacity under the limitation of secrecy outage probability.By jointly optimizing the power allocation between the artificial noise(AN)signal and secret signal of the transmitter,the closed-form expression of the secrecy outage probability constraint is given using analytical analysis method,and a lower bound of secrecy capacity under the influence of channel errors is derived.More importantly,a high-precision complex-valued neural network channel estimator is proposed to simulate the realistic communication scenario,effectively improve channel estimation accuracy,and reduce the impact of channel estimation errors on security performance.The numerical simulation results verify that the system can achieve almost the same security performance with that of the perfect channel case by using the complex-valued neural network channel estimator.(2)For the energy-constrained multi-task secure communication system,in order to improve secrecy energy efficiency,the jammer in the traditional cooperative jamming scheme is proposed to transmit public data to its service subscriber while enhancing the security performance of the secure transmission link.By optimizing resource management to maximize the secrecy rate of the system incorporating energy efficiency,the objective security rate function is proved to be a strictly concave function and the conditions for secure transmission of the communication system are clearly deduced.For the high noise or deep fading channel environment,the quality of wireless channel is too poor to meet the requirements of security and quality of service at the same time.This dissertation presents the corresponding dynamic adjustment strategy.(3)Considering attackers as multiple multi-antenna eavesdropping nodes,the existing cooperative jamming scheme is extended to a more general scenario in which the secondary user link not only harvests energy and transmits data,but also facilitates the secure transmission of confidential messages transmitted by the primary user.In view of the perfect channel and the norm-bounded CSI uncertainty,this dissertation studies the cooperative cognitive radio combined with the simultaneous wireless information and power transfer(SWIPT)communication technology.In this dissertation,an AN-aided secure beamforming scheme is proposed.The corresponding secrecy rate optimization problem is highly nonconvex,thus it is difficult to obtain the globally optimal transmit covariance matrix solution.In order to obtain the optimal or suboptimal transmit covariance matrix solution for the nonconvex security optimization problem,this dissertation adopts one-dimensional linear search method(i.e.,golden section algorithm)and matrix inequalities(e.g.,S-procedure,Shur complement,and first order Taylor series expansion)to transform the sophisticated natural logarithmic determinants and nonconvex matrix inequalities into the explicit solvable linear matrix inequalities(LMIs).Finally,the security optimization problem is conservatively approximated as the related semidefinite programming(SDP)problem.The solution of security problem is obtained by using convex optimization toolbox.(4)Considering the impact of multiple legitimate users on the security performance of communication system,this dissertation finally studies the adaptive security transmission of confidential data to multiple legitimate users.Under the Gaussian channel state information(CSI)uncertainty model,a robust AN combined with precoding scheme is proposed in the multi-input multi-output multi-eavesdropper(MIMOME)multibroadcast communication scenario.The transmitter multicasts the confidential signal to multiple multi-antenna receivers.For avoiding energy waste,idle users within the network coverage use energy harvesting devices to convert radio frequency energy into usable power and extend battery life.In this dissertation,the covariance matrices of the transmitted signals are optimized to maximize the secrecy rate of communication system under the joint constraints of secrecy outage,energy harvesting and transmit power.The proposed problem is abstracted into nonconvex mathematical expressions,and the mathematical process is extremely complicated.The original security optimization problem is converted into two convex subproblems by means of low SNR approximation and some other approximation methods(e.g.,Bernstein inequality and Sphere Bounding method).Then,the convex optimization toolbox is used to solve the problem,by which the approximate solution to the transmit covariance matrix is obtained.