Scholarlink

<p>In millimeter wave massive MIMO systems, hybrid beamforming and optimal antenna selection are essential for maximizing spectral efficiency and minimizing system latency. However, conventional beamforming approaches face challenges related to computational complexity and real-time performance. This paper proposes a novel deep learning framework using Long Short-Term Memory (LSTM) networks to streamline hybrid beamformer design and antenna selection. By framing the problem as a prediction and classification task, the LSTM processes the channel matrix to determine optimal subarray configurations and hybrid beamformers design, enhancing both efficiency and accuracy. To ensure robustness, the LSTM is trained on noisy channel matrices, addressing real-world variability in channel conditions. Experimental results demonstrate that the proposed LSTM-based approach achieves substantial improvements in spectral efficiency and computational complexity reduction compared to traditional beamforming methods, positioning it as a viable solution for high-performance millimeter wave MIMO systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A robust LSTM framework for hybrid beamforming and antenna selection in mmWave MIMO systems

  • 期刊论文

  • 发表日期：

    2026年5月4日

查看全文
  • Safa Awadallah,
  • Ali Hakan Ulusoy,
  • Mustafa Mulla,
  • Ahmet Rizaner

摘要

In millimeter wave massive MIMO systems, hybrid beamforming and optimal antenna selection are essential for maximizing spectral efficiency and minimizing system latency. However, conventional beamforming approaches face challenges related to computational complexity and real-time performance. This paper proposes a novel deep learning framework using Long Short-Term Memory (LSTM) networks to streamline hybrid beamformer design and antenna selection. By framing the problem as a prediction and classification task, the LSTM processes the channel matrix to determine optimal subarray configurations and hybrid beamformers design, enhancing both efficiency and accuracy. To ensure robustness, the LSTM is trained on noisy channel matrices, addressing real-world variability in channel conditions. Experimental results demonstrate that the proposed LSTM-based approach achieves substantial improvements in spectral efficiency and computational complexity reduction compared to traditional beamforming methods, positioning it as a viable solution for high-performance millimeter wave MIMO systems.