Scholarlink

<p>The performance of millimeter-wave (mmWave) massive MIMO systems relies significantly on the accuracy of channel state information to facilitate beam selection and ensure efficient communication. Channel estimation, however, presents a major challenge due to the dynamic nature of the environment and the high-dimensionality of mmWave channels. This paper introduces an innovative approach that enhances beamspace channel estimation by integrating Recurrent Neural Networks (RNNs) with the Gaussian mixture learned approximate message passing (GM-LAMP) method. The use of RNN layers enables the model to better capture both spatial and temporal dependencies within beamspace channels, thereby improving estimation accuracy, particularly in scenarios involving user mobility and fluctuating environments. RNNs efficiently monitor angular and path-related correlations by processing sequential data, which results in a more robust channel estimation procedure. The results of the simulation demonstrate that the proposed RNN-enhanced GM-LAMP framework outperforms the performance of current approaches regarding estimate accuracy and minimizes pilot overhead.</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

Recurrent Neural Networks for Enhanced Beamspace Channel Estimation in mmWave Massive MIMO Systems

  • 期刊论文

  • 发表日期：

    2025年2月1日

查看全文
  • Zaid Albataineh

摘要

The performance of millimeter-wave (mmWave) massive MIMO systems relies significantly on the accuracy of channel state information to facilitate beam selection and ensure efficient communication. Channel estimation, however, presents a major challenge due to the dynamic nature of the environment and the high-dimensionality of mmWave channels. This paper introduces an innovative approach that enhances beamspace channel estimation by integrating Recurrent Neural Networks (RNNs) with the Gaussian mixture learned approximate message passing (GM-LAMP) method. The use of RNN layers enables the model to better capture both spatial and temporal dependencies within beamspace channels, thereby improving estimation accuracy, particularly in scenarios involving user mobility and fluctuating environments. RNNs efficiently monitor angular and path-related correlations by processing sequential data, which results in a more robust channel estimation procedure. The results of the simulation demonstrate that the proposed RNN-enhanced GM-LAMP framework outperforms the performance of current approaches regarding estimate accuracy and minimizes pilot overhead.