Improved MMSE channel estimation in massive MIMO system with a method for the prediction of channel correlation matrix

Channel correlation matrix is an essential statistical measure that must be available to estimate minimum mean square error (MMSE) channel state information (CSI) for massive MIMO system and to carry out pilot reassignment optimization. Unless this information is available, the estimation error by the MMSE method will be high. In this work, we propose a new method to update channel correlation matrix based on last known CSI and temporal properties of the dynamic wireless channel. The predicted correlation matrix is then used to estimate MMSE CSI based on the user terminal pilot carrier. The results show that the proposed update approaches the ideal MMSE estimate when the temporal correlation is high and without a reduction in the capacity sum rate.