Intelligent Reflecting Surface Assisted NOMA with Heterogeneous Internal Secrecy Requirements

Abstract

The intelligent reflecting surface (IRS) technology can significantly enhance the flexibility of non-orthogonal multiple access (NOMA) by reconfiguring the channel gains according to the users' communication requirements. We propose to use IRS to enhance the internal secrecy among NOMA users in the multiple-input-single-output downlink networks. The objective is to minimize the total transmit power under heterogeneous secrecy constraints, by jointly optimizing the beamforming vectors and the IRS reflecting factors. To efficiently solve this non-convex problem, an alternating optimization based iterative algorithm is proposed by leveraging the successive convex approximation and the semi-definite relaxation techniques. Simulation results show that our proposed IRS-assisted NOMA can effectively guarantee the internal secrecy of users with significantly reduced transmit power compared to the convectional NOMA without IRS and the orthogonal multiple access with or without IRS.