An Exploration-Estimation Beamforming Scheme For 5GNR FDD Massive MIMO Communications

In real-life 5G massive multiple-input and multiple-output (MIMO) cellular systems, achieving high-performance beamforming heavily depends on accurate downlink channel state information (CSI) acquisition, which is particularly challenging in frequency division duplexing (FDD) systems due to a lack of channel reciprocity. To tackle this challenge with low communication overhead, current FDD systems adopt the limited feedback scheme to encode the downlink CSI estimated by the user into a few scalars and feed them back. However, this scheme hinders optimal beamforming vector acquisition at the base station due to the difficulty in CSI recovery. In this paper, we propose an online alternating exploration-estimation approach based on a comprehensive analysis of the limited feedback scheme to address this problem. By extracting information hidden in the feedback values, this approach can effectively estimate the optimal beamforming vector. Moreover, the beamforming exploration step in this approach can guide the limited feedback scheme to provide a more accurate CSI. Numerical results demonstrate the outstanding performance of the proposed approach in terms of optimal beamforming vector acquisition.