Bayesian Approach to Channel Interpolation in Massive MIMO Receiver

Abstract

We propose a new interpolation algorithm to enhance channel estimation accuracy in Massive Multiple Input, Multiple Output (MIMO) receiver. The algorithm employs a Bayesian approach to channel tap search and de-noise by taking into account the prior statistics of channel tap distribution. We validate our method using realistic state-of-the-art 5G simulations in a large number of non-line-of-sight scenarios of QuaDRiGa 2.0 channel. We show our method to outperform standard alternatives such as polynomial interpolation and time-domain windowing.