PDF and simulation code for the project "Spatial Correlation and Low Complexity Signal Processing Techniques in Massive MIMO Systems

One of the most promising technologies to meet society's current needs in (6G) wireless communications is the massive Multiple Input Multiple Output (M-MIMO) system. Initially, this work aimed to present the canonical concepts behind this technology in order to discover its functionalities and the open problems related to its envisaged implementation. As a fundamental complication of this system, pilot contamination has been investigated here; where a successive pilot decontamination method has been duly evaluated. Some drawbacks of this approach have been reported, whereby the use of multiple phases of pilot training has proved unattractive to manage in the event of pilot contamination. Motivated by the lack of studies associated with spatial correlation over multi-antenna channels in an M-MIMO scenario, the exponential antenna array correlation model combined with large-scale fading variations on the array was analyzed for favorable channel hardening and propagation effects; where, the normalized mean squared error (NMSE) was also deployed as an evaluation measure. Since the current interest lies in the use of different arrangements of antenna arrays, the performance evaluation of spatially correlated M-MIMO channels was carried out by deploying the Uniform Linear Array (ULA) and the Planar Array. uniform (UPA). Generally speaking, it has been found that the favorable propagation effect improves with spatial correlation, in which UPA guarantees better results. Since the favorable propagation is correctly related to the channel estimate, the pilot contamination was seen to be reduced by considering the use of the minimum mean square error (MMSE) estimator. Instead of these positive results, it was observed that channel hardening was poorly sustained under conditions of high spatial correlation. Considering these points, a trade-off between ensuring favorable propagation and channel hardening was discussed in order to support a more realistic network design based on the studied model. Another current research interest lies in reducing the complexity associated with linear signal processing techniques applicable to M-MIMO systems. With this in mind, the Kaczmarz (KA) algorithm has been implemented to solve combination / precoding problems following some recent directions found in the literature. In order to improve the convergence rate of this iterative algorithm,

Licence  GPL-3.0

Victor Croisfelt Rodrigues, "Spatial Correlation and Low Complexity Signal Processing Techniques in Massive MIMO Systems", Final year project, Universidade Estadual de Londrina, Londrina, Brazil, December 2018.