Secrecy energy efficiency optimization for AN-aided SWIPT system with power splitting receiver

Dear editor, Two emerging technologies of fifth generation wireless systems (5G), i.e., simultaneous wireless information and power transmission (SWIPT) and physical layer (PHY) security, are attracting considerable attention because they can reduce system energy consumption and can enhance network communication security, respectively [1–3]. These technologies have been jointly studied as secure SWIPT, and recent literatures have investigated the problems of secrecy rate maximization (SRM), harvested energy maximization, and transmission power minimization [4–6]. However, since SRM and harvested energy maximization have conflicting goals, secrecy energy efficiency (SEE) is considered to be the best tradeoff between them. A dearth of research in the SEE problem in secure SWIPT motivates this work. This study considers secure SWIPT with one base station (BS) and multiple users, where the message intended for a legitimate user (LU) should be kept secure from eavesdroppers (Eves). For improving communication security, artificial noise (AN) is transmitted together with the confidential message for LU, and all users (Eves and LU) implement a power splitting (PS) scheme for energy harvesting (EH). Due to the coupling of variables, the formulated SEE maximization (SEEM) problem is intractable. We propose to solve this using a low-complexity, two-stage SEEM algorithm that incorporates a slack variable and Dinkelbach method. The simulation results validate the efficiency of the proposed algorithm and show the advantage of AN in improving the SEE of the system. System model and problem formulation. We consider a downlink AN-aided secure SWIPT system, which consists of one BS with Nt > 1 antennas, one LU andK Eves, each of which is equipped with a single antenna. The perfect channel state information (CSI) of LU and Eves is assumed to be available at BS, and AN lies in the null space of LU’s channel [7]. The worst case, i.e., Eve uses the entire received signal for information decoding (ID) to achieve the maximum signal to interference plus noise ratio (SINR), is considered in this study. Thus, the PS ratio of the k-th Eve is ρk = 1, while that of LU is ρ (0 6 ρ 6 1). Let us denote the index set of Eves as K = {1, 2, . . . ,K}. Subsequently, the SINR for ID at LU and the k-th Eve can be expressed respectively as