Channel Magnifying Limited Feedback Technique for FDD Massive MIMO Systems

Abstract

A limited feedback system, so-called, channel magnifying (CM) is proposed for a downlink (DL) frequency-division duplex (FDD) massive multiple-input multiple-output (MIMO) system. Although massive MIMO system has received significant research interest as a key technology for beyond 4G wireless communications systems, it has a number of issues that needs to be technically addressed. Among such issues is the difficulty of acquiring channel state information at transmitter for an FDD massive MIMO system which cannot exploit channel reciprocity as in a time-division duplex system. The proposed CM technique makes it possible to support a few user equipments in DL FDD massive MIMO system by finding a balance between spatial resources and channel quantization error (CQE). By choosing a subchannel with low CQE, CM can secure multiplexing gain at high SNR based on a fixed size codebook. Two types of subchannel indicator alignment (SIA) schemes are introduced for efficient interference nulling for the proposed CM technique. Specifically, we discuss how to maximize the sumrate of CM through genie added SIA and minimum CQE based SIA. Simulation results show that the sum rate of the proposed CM has a higher multiplexing gain than that of random vector quantization, especially when the number of transmit antennas is sufficiently large.