MISO Broadcasting for SWIPT with Batteryless Receivers and Battery-Constrained Self-Sustainable IRS

This paper considers a multiple-input single-output broadcasting system for simultaneous wireless information and power transfer (SWIPT) with batteryless user terminals (UT) which is aided by a self-sustainable intelligent reflecting surface (IRS). The UTs and the battery-powered IRS employ the time switching (TS) or the power splitting (PS) scheme. The various combinations of these SWIPT protocols result in different system operation modes. Considering a nonlinear energy harvesting (EH) model, we aim at jointly optimizing in each mode the transmit and reflect beamforming vectors, along with the TS or/and PS ratios, such that the transmit sum-power is minimized subject to the quality-of-service requirements of the UTs, the EH requirements of the IRS, and its battery's capacity. We formulate, whenever required, corresponding multi-objective optimization problems and convert them into equivalent single-objective optimization problems using the epsilon-constraint method. We develop iterative algorithms under the alternating minimization or block coordinate descent frameworks that utilize the semi-definite relaxation, inner approximation, and successive convex approximation methods, along with low-complexity heuristics or/and line search, to efficiently tackle the challenging non-convex optimization problems under study. Numerical simulation results highlight the performance gains of the proposed resource allocation strategies over baseline techniques and provide valuable insights.