Performance Evaluation of a Low Complexity Hybrid Beamforming for 5G mmWave MIMO Systems
No Thumbnail Available
Date
2025-11
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
The demand for higher network capacity is drastically increasing and the effort to exploit new capacity boosting techniques are also growing simultaneously. Fifth-Generations networks have found millimeter wave signals, from 10mm to 1 mm wavelengths, to be the ideal frequency ranges to support the growing higher data rates requirement which reaches up to 10 Gbit/s with a vast bandwidth.
In massive MIMO systems, dependency on traditional beamforming methodologies is no longer valid due to issues such as increased computational burden, scalability issues, cost, and power requirements. For large-scale systems that are limited by power and cost the use of fully digital beamforming is not practical anymore because the baseband can only accommodate a certain number of analog to digital converters (ADCs) and signal mixers. The analog beamforming technique, however, is limited by flexibility and scalability issues, and unsuitable for MIMO systems.
To mitigate the mmWave signal attenuation and achieve higher data speeds as well as improved spectral efficiency, deployment of massive MIMO and hybrid analog-digital beamformers is now viewed as a feasible solution. In this thesis, a performance evaluation and complexity analysis has been carried out for the low complexity hybrid beamforming technique called the PE-AltMin algorithm. In particular, the spectral efficiency and computational complexity evaluation of this algorithm is studied in contrast with the conventional fully digital optimal beamformer and the MO-AltMin algorithm for the fully connected hybrid beamformer mapping. This thesis paper has shown that PE-AltMin algorithm is computationally less complex than MO-AltMin and fully-digital beamforming techniques. The PE-AltMin algorithm provides favorable tradeoff for better computational resources utilization where hardware simplicity matters, and real-time deployment are prioritized. MO-AltMin algorithm is preferable when computational resources are not the challenge and performance margins matter the most.
Description
Keywords
mmWave, Massive MIMO, Hybrid Beamforming, Alternating Minimization, Phase Extraction, Manifold optimization, Conjugated Gradient Descent (CGC), Riemannian Conjugate Gradient (RCG), Low complexity