STAR-RIS Aided SWIPT Systems With Non-Linear Energy Harvesting Model

Simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) brings a great potential due to its full-angle coverage characteristic. In this work, a multi-user multiple-input single-output simultaneous wireless information and power transfer network aided by a coupled phase-shift STAR-RIS is proposed, where each co-located receiver having the capabilities of decoding signals by an information decoder and harvesting energy by an energy harvester, is located in the reflecting area or transmitting area. In addition, a practical non-linear energy harvesting (EH) model is introduced in the proposed scheme for avoiding resource allocation mismatches caused by linear EH model. In this case, a sum rate maximization problem is formulated, which is non-convex. To tackle the non-convexity brought by coupled phase-shift constraint, we propose two efficient algorithms, namely penalty dual decomposition (PDD)-based and consensus alternating direction method of multipliers (CADMM)-based algorithms. In addition, after transforming the problem by Lagrangian dual transform and quadratic transform methods, the block coordinate descent method is introduced to divide the formulated problem into several subproblems. After solving the formulated problem based on the proposed efficient methods, our numerical results show some insights: 1) The proposed scheme with STAR-RIS is superior to the scheme with the conventional RIS; 2) The CADMM method enjoys less computation time as compared to the PDD method, and has a similar sum rate value when the number of elements is large; 3) The PDD method is more suitable as compared to the CADMM method when the number of elements is small because of its prominent sum rate.