On Channel Estimation for Massive MIMO With Pilot Contamination

This letter proposes a simple channel estimator for a multi-cell massive MIMO system with pilot contamination. Under moderate or strong pilot contamination, the existing least square (LS) channel estimator suffers from substantial performance degradation while the existing minimum mean square error (MMSE) channel estimator offers much improved performance over the LS estimator. However, the existing MMSE estimator assumes perfect knowledge of cross-cell large-scale channel coefficients and such an assumption is practically unjustifiable. The proposed estimator overcomes the issues of both existing estimators and provides estimation performance close to that of the ideal MMSE estimator. An analytical mean square error (MSE) expression is also derived for the proposed estimator.