Massive multiuser MIMO downlink with low-resolution converters

In this review paper, we analyze the downlink of a massive multiuser multiple-input multiple-output system in which the base station is equipped with low-resolution digital-to-analog converters (DACs). Using Bussgang’s theorem, we characterize the sum-rate achievable with a Gaussian codebook and scaled nearestneighbor decoding at the user equipments (UE). For the case of 1bit DACs, we show how to evaluate the sum-rate using Van Vleck’s arcsine law. For the case of multi-bit DACs, for which the sum-rate cannot be expressed in closed-form, we present two approximations. The first one, which is obtained by ignoring the overload (or clipping) distortion caused by the DACs, turns out to be accurate provided that one can adapt the dynamic range of the quantizer to the received-signal strength so as to avoid clipping. The second approximation, which is obtained by modeling the distortion noise as a white process, both in time and space, is accurate whenever the resolution of the DACs is sufficiently high and when the oversampling ratio is small. We conclude the paper by discussing extensions to orthogonal frequency-division multiplexing systems; we also touch upon the problem of out-of-band emissions in lowprecision-DAC architectures.