Projection-based Differential Feedback for FDD Massive MIMO Systems

Abstract

Channel state information at the transmitter (CSIT) plays a key role in achieving potential gain of massive multiple-input–multiple-output (MIMO) systems. In frequency-division duplex (FDD) systems, CSIT can be obtained through feedback from the receiver. Conventional limited feedback schemes, which have been designed for small-scale MIMO systems, suffer from a prohibitive amount of feedback requirement and encoding complexity when the number of transmit antennas becomes massive. In this paper, a projection-based differential feedback (PBDF) protocol is proposed for FDD massive MIMO systems. In the PBDF framework, the difference between original and predicted vectors is projected, and quantization is performed in a smaller dimensional subspace. With an appropriate projection exploiting spatial and temporal correlation of massive MIMO channels, the feedback amount and encoding complexity can be significantly reduced. The simulation results show that the proposed scheme achieves a large portion of potential throughput gain of massive MIMO systems with a small amount of channel feedback.