Bayesian Learning-Based Doubly-Selective Sparse Channel Estimation for Millimeter Wave Hybrid MIMO-FBMC-OQAM Systems

We design and analyse filter bank multicarrier (FBMC) offset quadrature amplitude modulation (OQAM)-based millimeter wave (mmWave) hybrid multiple-input multiple-output (MIMO) systems. Furthermore, a novel channel estimation model is conceived for quasi-static mmWave hybrid MIMO-FBMC-OQAM (mmH-MFO) systems that reconfigures the radio-frequency (RF) circuitry during the transmission of zero symbols. Subsequently, a Bayesian learning (BL) technique is proposed for sparse channel estimation, which relies on multiple measurement vectors combined with selective subcarrier grouping for enhanced estimation. Additionally, an online BL based Kalman filter (OBL-KF) is designed for sparse channel tracking in doubly-selective mmH-MFO systems. Then the Bayesian Cramér-Rao lower bounds (BCRLBs) are derived for characterizing the performance of the proposed frequency-selective and doubly-selective channel estimation techniques. Finally, a limited feedback based algorithm relying on beamspace channel estimates is proposed for hybrid precoder/combiner design. The accuracy of our analytical results is confirmed by our simulation results.