RSMA-Enhanced Secure Transmission in IRS-Assisted Networks Against Internal and External Eavesdroppers

In this letter, we consider an intelligent reflecting surface (IRS)-assisted network consisting of two legitimate users and multiple eavesdroppers. For realizing the secure transmission of confidential information, the rate-splitting multiple access (RSMA) technique is investigated to against the internal and external eavesdropping. Based on the design principle of RSMA, the confidential information can be placed into the private message and concealed within the common message to achieve secure transmission. Therefore, we aim to minimize the transmit power of private message by jointly designing transmit beamforming of base station (BS) and IRS phase shifts. Due to the non-convexity of the optimization problem with multiple high coupling variables, the problem can be divided into two subproblems and an alternating optimization (AO) framework is adopted. To deal with the subproblems, an algorithm based on semidefinite relaxation (SDR) and successive convex approximate (SCA) is proposed. Simulation results demonstrate that the proposed RSMA-enhanced scheme can improve the security of confidential information in the scenario of existing both internal and external eavesdroppers, while outperforming other schemes.