MIMO-FBMC Channel Estimation with Limited, and Imperfect Knowledge of Channel Correlations

This paper presents and analyses the performance of training-based least squares (LS) and minimum mean square error (MMSE) channel estimation schemes for multiple input multiple output (MIMO) filter bank multicarrier (FBMC) systems based on the offset quadrature amplitude modulation (OQAM) in the presence of limited, and imperfect knowledge of the channel correlations. First, a linear MMSE (LMMSE) technique for MIMO-FBMC channel estimation, which require a priori knowledge of channel correlation matrix, is examined by utilizing the second-order statistical properties of the intrinsic interference in FBMC systems. A biased LS (BLS) and relaxed LMMSE (RLMMSE) MIMO-FBMC channel estimation schemes, which require prior knowledge of the trace of the channel correlation matrix, are proposed. The LS-BLS and LS-RLMMSE schemes for MIMO-FBMC channel estimation are investigated in the presence of imperfect knowledge of the channel correlations. The mean square error is derived for the proposed schemes by exploiting statistical properties of the intrinsic interference. Simulation results show that the proposed schemes present an excellent trade-off between the achieved performance and required a priori knowledge of the channel correlations.