Efficient Channel Estimation and Extrapolation for Pattern Reconfigurable Massive MIMO with Low Pilot Signaling Overheads

Reconfigurable antennas have excellent dynamic adaptability to alter their operational state in response to environmental changes, and is considered as one potential technology towards future communication systems. However, acquiring accurate channel state information (CSI) for all radiation patterns with low pilot overheads imposes a significant challenge in pattern reconfigurable MIMO (PR-MIMO) communication systems. To address this issue, in this paper we propose a novel two-stage channel estimation approach based on antenna grouping (AG) to obtain the CSIs for all radiation patterns efficiently. In the first stage, all antennas at the transmitter employ the same radiation pattern, while in the second stage, the antennas at the transmitter are divided into groups according to the number of radiation patterns, where antennas in different groups employ different radiation patterns, while antennas within the same group employ the same radiation pattern. When the exact number of channel paths is known, a closed-form channel extrapolation algorithm and a singular value decomposition (SVD)-based channel extrapolation algorithm are proposed, depending on the value of the channel paths and whether the angle information is known. Extensive simulation results illustrate that the proposed algorithms can accurately extrapolate the CSI of all radiation patterns, with dramatically reduced pilot overheads compared to the conventional channel estimation methods.