Massive MIMO for SWIPT: A Measurement-Based Study of Precoding

Abstract

Massive multiple-input multiple-output transfer (MaMIMO) has been shown to be a viable technology for improving spectrum and energy efficiency of future wireless generations, both using theoretical and measurement-based approaches. While the benefits of Massive MIMO for simultaneous wireless information and power transfer (SWIPT) have been established in theory, no practical evaluations have been reported. In this paper, we evaluate the power transfer efficiency of various established precoding algorithms, using random data signals and two methods to deal with the necessity of channel state information. A first method relies on uplink pilots transmitted by the energy harvesting receiver, while for the second method we use channel beacons to avoid that every sensor has to consume power to transmit pilots. We experimentally evaluate the efficiency of both approaches as function of the number of antennas and the precoder used and show that increasing the amount of antennas from 2 to 64 while fixing total transmit power results in a 10 fold increase in rectifier output voltage. When using a channel beacon, we show that receive power can be sacrificed to construct broader beams, which can be shared by different wirelessly powered sensors. All measurements report rectifier output voltage, as measured in our facility.