Performance evaluation of massive MIMO Systems for future wireless access systems

Gahadza, Mutsawashe

Performance evaluation of massive MIMO Systems for future wireless access systems

Master Thesis

2019

Abstract

The vision for 5G is predicated on three main cornerstones. These are massive Machine Type Communication (mMTC) technologies, Ultlra Reliable Low Latency Communications (uRLLC) and enhanced Mobile Broadband (eMBB). In order to achieve the high capacity needed for enhanced Mobile Broadband (eMBB) to be a reality, a number of technologies have been proposed in various research forums. These include use of spectrum bands like 26GHz which are called the mmWave spectra. Although the use of mmWave spectra brings in a lot of capacity because of increased bandwidth, the signal attenuates quickly as it suffers a lot of diffraction losses at such high frequencies. In order to mitigate this the use of massive MIMO technology has been proposed. Massive MIMO improves both spectral efficiency and energy efficiency and is therefore also proposed for spectra below 6GHz. This study focuses on assessing the potential of massive MIMO through evaluation of linear precoding and receive combining methods that may be applicable to massive MIMO. Linear signal detection and precoding for MIMO is generally not optimal. Optimal methods such as Maximal Likelihood (ML) signal processing methods have high computational complexity such that their practical implementation is difficult. The complexity for ML is O(MN ) for an M−ary modulated signal and N antennas. This is compared to a linear signal processing method called Zero Forcing (ZF) with a complexity of the order of O(N3 ). Assessing the performance of linear signal processing methods is therefore invaluable for the success of massive MIMO in general and 5G in particular. Simulations to evaluate spectral efficiency for massive MIMO were done in MATLAB. Linear and sub-optimal signal processing methods like minimum mean square error (MMSE), zero forcing (ZF), regularized zero forcing (RZF) and maximal ratio combining (MRC) detection and precoding algorithms with relatively less complexity were evaluated. The spectral efficiency (SE) of these signal processing methods were evaluated through a Monte Carlo simulation method in a massive MIMO single base station cell, a 16 cell grid network and a 64 cell grid network. SE values of up to 200 kbps/Hz/cell were obtained with 100 antenna elements and 10 users per cell. The effect of pilot reuse factor for both detection and precoding signal processing systems was also evaluated. A pilot reuse factor of 8 seemed optimal for 64 cell grid network modeled. The overall results obtained from this study show that the Spectral Efficiency (SE) improves as the number of antenna elements to users ratio increased. MMSE and RZF had the best performance under all simulation conditions while for Maximal Ratio Combining (MR) a much larger number of antenna elements was needed in order to approach the performance of MMSE and RZF. An evaluation of the effect of the iii dominant propagation channel conditions was also done by evaluating the spectral efficiency performance of the four detection methods in correlated and uncorrelated channels. Lastly the effect of pilot contamination was investigated. The results showed that an optimal value that maximizes the obtainable spectral efficiency for a massive MIMO network can be obtained.

Keywords

Electrical Engineering

Reference:

Collections

Masters

Full item page