Hybrid Polarized Beamforming Optimization of Millimeter Wave Multi-User Massive MIMO Communication Systems

An appropriate hybrid beamforming scheme is essential for millimeter wave (mmWave) multi-user massive Multiple-Input-Multiple-Output (MU-mMIMO) communication architectures, as it ensures both the advantages of digital processing and the efficacy of analog technology. To further explore the channel spectrum on the polarization domain, polarization selectivity is introduced in this work to improve the performance of the beamforming design in mmWave MU-mMIMO systems. By introducing single-port polarization reconfigurable antenna units to the mMIMO antenna arrays, we propose a novel hybrid polarized beamforming structure that can flexibly adjust the polarization distribution of the array. Based on the structure, this paper models the polarization distribution of the subarrays as selection of mask matrices and proposes several optimization schemes to enhance the hybrid polarized beamforming performance, including random combining, conventional genetic algorithm (GA), and optimal perturbation genetic algorithm (OP-GA). In this paper, we enable the Saleh-Velenzuela (S-V) millimeter wave channel model to be applied to the analysis by extending the multipath component (MPC) polarization distribution following the 3GPP and WINNER channel models. Performance analysis on the system sum rate shows that the proposed design significantly outperforms the fixed-polarized system under the same channel distributions. This paper underscores the importance of polarization optimization and provides a reference for enhancing the performance of mmWave MU-mMIMO systems.