Multi-Beam Beamforming in 5G Massive MIMO Networks: Normalization Techniques and Algorithmic Performance Evaluation

Enahoro, Sunday, Ekpo, Sunday Cookey ORCID: https://orcid.org/0000-0001-9219-3759, Ji, Helen ORCID: https://orcid.org/0000-0001-7955-2999, Chow, Kin Kee ORCID: https://orcid.org/0000-0003-0589-6208, Karimian, Noushin, Uko, Mfonobong, Elias, Fanuel, Unnikrishnan, Rahul, Olasunkanmi, Nurudeen Kolawole and Alabi, Stephen (2025) Multi-Beam Beamforming in 5G Massive MIMO Networks: Normalization Techniques and Algorithmic Performance Evaluation. IEEE Access. pp. 1-21. ISSN 2169-3536

Preview

Accepted Version Available under License Creative Commons Attribution. Download (1MB) | Preview

Abstract

The challenge of optimally allocating transmit power among multiple beams remains a central problem in multi-beam beamforming for massive MIMO networks, especially as user density and interference become increasingly dynamic in next-generation wireless deployments. Conventional normalization techniques, such as Global and Per-Beam Normalization, are unable to adapt to varying channel conditions and user demands, often resulting in inefficient power distribution and limited system performance. To address this, we propose Power Redistribution Normalization (PRN), a novel approach that dynamically reallocates transmit power across multiple beams in real time, optimizing both desired signal peaks and interference nulls. The PRN framework is integrated with diverse beamforming algorithms—including Least Mean Square (LMS), Minimum Variance Distortionless Response (MVDR), Recursive Least Squares (RLS), Zero Forcing (ZF), and Sample Matrix Inversion (SMI)—and is evaluated under realistic antenna configurations, user densities, and interference scenarios. Results demonstrate that PRN-MVDR achieves the highest spectral efficiency (~25 bps/Hz), SINR (~8.7 dB), and energy efficiency (~25 bps/Hz/W) compared to existing normalization methods. Moreover, PRN-LMS provides comparable performance with significantly lower computational complexity (~0.4 ms latency), making it attractive for real-time and hardware-constrained applications. The proposed PRN method is thus well-suited for dense urban 5G/6G networks and large-scale IoT deployments, where robust power allocation and interference management are critical to system capacity and reliability.