Learning Based Hybrid Beamforming for Millimeter Wave Multi-User MIMO Systems
This work addresses efficiency and robustness issues in wireless communication systems, representing an incremental improvement over conventional methods.
The authors tackled the problem of high complexity and channel state information reliance in hybrid beamforming for millimeter wave multi-user MIMO systems by proposing an extreme learning machine framework, which achieved higher system sum-rate and reduced computation time compared to existing methods.
Hybrid beamforming (HBF) design is a crucial stage in millimeter wave (mmWave) multi-user multi-input multi-output (MU-MIMO) systems. However, conventional HBF methods are still with high complexity and strongly rely on the quality of channel state information. We propose an extreme learning machine (ELM) framework to jointly optimize transmitting and receiving beamformers. Specifically, to provide accurate labels for training, we first propose an factional-programming and majorization-minimization based HBF method (FP-MM-HBF). Then, an ELM based HBF (ELM-HBF) framework is proposed to increase the robustness of beamformers. Both FP-MM-HBF and ELM-HBF can provide higher system sum-rate compared with existing methods. Moreover, ELM-HBF cannot only provide robust HBF performance, but also consume very short computation time.