Data-aided iterative reweighted LMMSE channel estimation for MIMO OFDM

In this paper, we investigate data-aided reweighted Linear Minimum Mean Square Error (LMMSE) iterative channel estimation in a Multi Input Multi Output (MIMO) OFDM communication system. The estimation is BEM (Basis Expansion Model) based. The channel estimator with every iteration improves estimates through the use of tentative data symbols and corresponding adaptive weights. The weights are computed based on residual powers via non-linear mapping. The initial channel estimates are obtained using the conventional Least Squares (LS) BEM pilot-based channel estimator. The performance of the reweighted iterative estimator is compared with that of the conventional unweighted LMMSE estimator. Numerical results for a downlink LTE scenario show that the reweighted channel estimation provides a significant improvement in the channel estimation and detection performance over the unweighted iterative estimation.