Robust Regularised Channel Inversion for Massive MIMO Unlicensed

In this paper, we propose a robust regularised channel inversion (RRCI) approach to improve the downlink performance of massive multiple-input multiple-output unlicensed (mMIMO-U) under imperfect channel state information (CSI). We first derive the optimal parameter $alpha$ for regularised channel inversion (RCI). We then compare the beamforming performance among our proposed RRCI approach, the conventional zero-forcing (ZF) beamforming and the conventional RCI in terms of cellular downlink capacity. We show that our proposed RRCI scheme outperforms the ZF beamforming and the conventional RCI. Moreover, our proposed RRCI can achieve a higher downlink sum rate when serving more cellular users.