Advanced Methods for Minimizing Pilot Overhead in Sparse Channel Estimation for Massive MIMO Systems

Abstract

There are various difficulties in doing downlink beam formation with the current high gain frequency division duplex (FDD) Massive multiple-input, multiple-output (MIMO) systems. In particular, downlink beam formation necessitates a channel estimate that often necessitates extensive training and feedback overhead, scaling with the base station's (BS) number of antennas. Assuring overhead reduction that is proportionate to the channel's level of sparsity, we use compressive sensing (CS) methods to properly estimate the channel. Through the suggested dictionary design, which is more adaptable, reliable, and capable of estimating the cell properties, the sparse virtual channel representation is created. We pay particular attention to huge MIMO- CS algorithms to determine which one works best with the suggested dictionary architecture. Orthogonal Frequency-Division Multiplexing (OFDM) systems that exhibit greater resistance to multipath fading. The fundamental pursuit bound is approached by greedy solutions with less complexity and a correspondingly shorter training duration, according to numerical data. With respect of performance complexity trade-off, the greedy algorithm with the best results is the normalised hard thresholding pursuit (NHTP) method.