Delay Reduction of Detection Algorithms for 5G Massive MIMO System

Abstract

Multiple antenna technologies like Multiple-InputMultiple-Output (MIMO) and beamforming will thus play animportant role in defining 5G system architectures. In massiveMIMO there is a huge number of antenna elements, so there isa need to estimate large channel matrix which introduces muchlatency. The ultra-high latency and high computation complexityof massive MIMO matrices from 16 to 256 dimensions is thevital bottleneck to realizing latency for channel estimation andMIMO detection. This paper introduces a mechanism to reducethe high computational complexity that causes huge latency. Fouralgorithms are evaluated to measure their performance. Thesealgorithms are Gauss-Jordan Elimination, Gaussian Elimination,RQ Decomposition and LU Decomposition. MATLAB simulationused to analyze the applied mathematical models. After thatmeasured the BER, delay for each algorithm and evaluate thecapacity and throughput, by way, found that the GaussianElimination has better delay about 49 percent when RQ Decomposition higherabout 95 percent while LU Decomposition highest about 98 percent comparedby Gauss-Jordan Elimination. In addition the result show theperformance of capacity and throughput for various modulationand coding rate, while the deliverables average capacity about10 M bit and affected by the situation of the channel, LU hasthe best performance than others.