Research On Key Technologies Of Physical Layer Security In Reconfigurable Intelligent Surface Aided NOMA System

The reconfigurable intelligent surface(RIS),a kind of passive device,can reflect signal from the base station(BS)and adjust the phase and amplitude of the reflected signal.As a revolutionary new technology,RIS can intelligently reconfigure the propagation environment of radio electromagnetic waves,which breaks through the limitation of uncontrollable traditional wireless communication channel.Non-orthogonal multiple access(NOMA)implements multi-user multiplexing in the power domain,which allows different users to share same resources in the time and frequency domain,and the receiver uses serial interference cancellation(SIC)technology to realize multi-user detection.Therefore,it greatly improves the access quantity and spectrum efficiency of communication system.In recent years,the intrinsic integration of the RIS and NOMA is proposed to improve the performance of wireless communication system.Most researchers focus on the design of algorithms to improve the energy efficiency,spectrum efficiency and transmission rate in the RIS-assisted NOMA system,while there are few research results on improving the physical layer security(PLS)performance of the system.Motivated by this situation,the beamforming algorithm and power allocation schemes are presented to improve the PLS of the system in the thesis.Firstly,the single input single output(SISO)NOMA system is studied for comparing the anti-eavesdropping performance of amplify and forward(AF)relay and RIS.The AF relay and RIS assisted transmission models are established respectively.In the case that the near user is untrusted,the performance of anti-internal eavesdropping with AF relay and RIS are compared.A closed numerical expression is obtained by theoretical analysis.When the number of reflecting elements of RIS is greater than a certain value,the PLS performance of RIS is better than that of AF relay at any signal to noise ratio(SNR).The results are verified by numerical simulation.Secondly,for the anti-eavesdropping problem of multiple input single output(MIMO)NOMA system,the system models with internal eavesdropping and external eavesdropping are studied.In the absence of line of sight(LOS)signals between the BS and the users,the order of SIC is switched by modifying the power allocation of the traditional NOMA system to prevent near user from eavesdropping far user.An alternate iterative algorithm is proposed to realize the joint beamforming for enhancing the channel condition of far user,and a dynamic power allocation scheme is designed to maximize the secrecy capacity of the system.On the basis,artificial noise(AN)is introduced to disturb external eavesdroppers.A noise beamforming algorithm based on Schmidt orthogonalization is proposed to allocate AN into the null space of the channel for NOMA users,and a joint dynamic power allocation scheme is presented to improve the PLS performance of the system.Simulation results show that increasing the number of reflecting elements of RIS or transmit antennas of BS have a positive impact on the system secrecy performance where increasing the number of reflecting elements of RIS can bring more secrecy gain than that of transmit antennas of BS.Finally,in the case that LOS signals are included between BS and users,a beamforming algorithm is proposed to against multiple external eavesdroppers.Based on the existing beamforming algorithm which refer to phase cancellation,a weighted beamforming algorithm based on channel gain is presented.The reflected signal of RIS and the direct signal of BS are canceled out at the external eavesdroppers and simultaneously increased at the NOMA users.Theoretical analysis and numerical simulation show that the PLS performance of the proposed algorithm is superior than the existing algorithm.