Uplink performance of massive MIMO subject to delayed CSIT and anticipated channel prediction

We consider a realistic cellular multi-user uplink channel with an excess of base station (BS) antennas, where the number of BS antennas and the number of users per cell increase at the same rate. Specifically, we investigate the negative impact of delayed channel state information at the transmitter (CSIT), when a minimum-mean-square-error (MMSE) detector is applied. Nevertheless, channel prediction is used for overcoming delayed CSIT degradation. We provide the asymptotic signal-to-interference-plus-noise ratios (SINRs) that demonstrate not only their dependence on delayed and predicted CSIT, but also the outperformance of MMSE against maximal ratio combiner (MRC). The deterministic nature of the results renders them easily computed, while simulations show their accuracy even for practical system dimensions.