Blind SINR Estimation for Massive MIMO Systems

The acquisition of Channel State Information (CSI) and the related Signal-to-Interference-plus-Noise Ratio (SINR) is a key challenge in the context of massive Multiple-Input Multiple-Output (MIMO) systems. When the number of antennas and users grow large, blind estimation methods are an attractive alternative to alleviate the burden of training overheads. These methods typically rely on the estimated covariance matrix of the received signals, which is inconsistent if the number of samples is comparable to the array size. In this setup, we propose a method to improve SINR accuracy using unbiased estimators for the double asymptotic regime of large numbers of antennas and samples. The benefits of the improved SINR estimates are shown in the context of adaptive modulation, precoding and power allocation, with substantial gains over state-of-the-art estimators.