Optimal Uplink Channel Estimation Algorithm for OFDM Based MmWave Massive MIMO Systems

Practical channel estimation algorithms for millimeter wave massive multiple-input multiple-output (mmWave massive MIMO) systems with huge number of antennas at the base station must achieve high spectrum efficiency. Furthermore, the uplink channel estimation becomes very challenging since the required pilot overhead used for channel estimation and feedback can be prohibitively large. In conventional channel estimation algorithms, the channel model is approximated using a virtual channel model with quantized angles of arrival/departure (AoA/AoD). In this paper, we consider the continually distributed AoA/AoD, we show that the mmWave massive MIMO channels share common cosparsity properties, and we propose a structured analysis compressive sensing (SACS) based algorithm which exploits those common cosparsity properties for channel estimation with low overhead. Simulation results show that our proposal can accurately estimate the channel with low overhead, and is capable of attaining the optimal Cramer-Rao Lower Bound (CRLB).