Semi-Blind Channel Estimation in Massive MIMO by Mutual Information Rate as Contrast Function

Abstract:
A massive multiple-input multiple-output system is considered. In order to improve channel estimation, we propose a semi-blind data-aided method based on complex independent component analysis (ICA) and using the mutual information rate as the contrast function. We do not require the powers of the interfering users to differ from the powers of the users of interest. The noise is first suppressed by subspace projection, then the uplink data symbols including the signal of interest and the interfering signals are extracted by ICA. Finally, the detected data symbols are used as pseudo-pilots to estimate the channel matrix of the users in the cell of interest. Simulation results confirm that the proposed method outperforms previous blind channel estimation methods.