Wireless Information and Power Transfer in Relay-Assisted Downlink Massive MIMO

The feasibility of utilizing a relay-assisted downlink massive multiple-input multiple-output (MIMO) to boost the performance of simultaneous wireless information and power transfer (SWIPT) is investigated. The key idea is to simultaneously transmit power and information toward the direct/relayed-users by leveraging the excess degrees-of-freedom offered by massive antenna arrays at the base-station (BS). The performance limits are established by deriving the harvested energies, achievable rates, and energy-rate trade-offs in the presence of imperfectly estimated/partial channel state information (CSI). These performance metrics are derived for a generalized SWIPT model, which facilitates analysis for both time-switching (TS) and power-splitting (PS) protocols. Two linear precoders based on zero-forcing and maximal ratio transmission are used at the massive MIMO BS, and a practically viable non-linear energy harvesting model is adopted for the direct/relayed-users. Joint detrimental implications of imperfect CSI at the BS, limited/statistical CSI at the users, and non-linear characteristics of energy harvesters are analytically quantified. In order to jointly guarantee user-fairness in terms of both energy and rate, a max-min optimal transmit power control policy is proposed. Then, the max-min fairness optimal energy-rate trade-offs are quantified for non-linear massive MIMO SWIPT. Our results reveal that the proposed relay-assisted downlink massive MIMO can be exploited to boost the achievable energy-rate trade-off by effectively reducing the end-to-end path-loss via shorter hop distances rendered by intermediate amplify-and-forward relays.