Spectral Efficiency Analysis in Downlink Massive MIMO System for Perfect CSI with Precoding

TL;DR

This paper derives and compares theoretical and simulated spectral efficiency bounds for downlink massive MIMO systems with perfect CSI, analyzing the impact of various parameters and algorithms.

Contribution

It provides a mathematical derivation of lower bound spectral efficiency for ZF and MMSE precoding, validated by simulations under practical channel conditions.

Findings

Theoretical bounds are 1 to 1.5 bits less than simulation results.

Algorithms-1 and ZF outperform MMSE in spectral efficiency.

Large scale fading significantly impacts spectral efficiency.

Abstract

In this paper we first derived the mathematical expression for lower bound spectral efficiency (SE) calculation for zero-force (ZF), and minimum mean square error (MMSE). Secondly, we calculated the simulation SE with three algorithms for ZF and MMSE precoding. We compared the simulation and theoretical results and found that the theoretical results are 1 to 1.5 bits less than the simulation values which implied that the theoretical lower bounds are actually the lower bounds. To achieve the maximum spectral efficiency in downlink massive MIMO systems we assumed perfect CSI, ZF and MMSE precoding in this paper. We also considered that the channel has the characteristics of small and large scale fading (SSF and LSF) as the model is like a practical. We investigated the effect of different SNR, base station (M) and radius (R) of the cell on spectral efficiency for simulation and theoretical results. We also evaluated the performance of SE of each algorithms and precoding schemes for different configurations. From the results we have observed that algorithms-1 and ZF outperform other algorithms and MMSE. From our investigation we noticed that the LSF parameters are the most dominated factor in SE in massive MIMO systems.