Scholarlink

<p>The fifth Generation (5G) communication device demands large connectivity with low delay and high data rates. One of the innovations to encounter these demands is millimeter Wave (mmwave) massive Multi-Input Multi-Output (MIMO). The hybrid beamforming and estimation of channel are taken for the multi-user mmwave MIMO device. In the modern days, various strategies have been recommended to model the hybrid beamforming and estimation of channels in mmwave MIMO devices. The conventional approaches concentrated on narrow-band channels. But, to efficiently employ the mmwave MIMO frameworks with higher bandwidth, there are modern tasks toward implementing the broadband hybrid beamforming mechanisms. The primary complexity in the hybrid beamforming is modeling a general beamformer, which is connected over the entire subcarriers while the weights of the digital beamformer are required to correspond to the subcarrier. In addition, the estimation of channels is complex because of the restricted Radio Frequency (RF) chains at the transceivers. Though numerous compressing sensing mechanisms have been implemented to rectify this issue by adopting sparsity frameworks, inherent channel sparsity, and practical works including beam squint, and power leakage can still create the original channel attributes that vary from desired techniques and lead to degradation in the functionality. Thus, in this work, a deep learning-assisted channel estimation technique with hybrid beamforming is implemented in the mmwave massive MIMO. At first, the necessary data is created for different scenarios to perform the channel estimation and hybrid beamforming process. Further, the Adaptive Convolutional Autoencoder (ACA) technique is supported to determine the channels in the mmwave massive MIMO device. Here, to increase the network efficiency and also provide better data transmission, the recommended ACA technique parameters are optimally tuned by the novel Hybrid Clouded Leopard and Walrus Optimization (HCLWO). Subsequently, the channel estimation and beam forming vectors are optimized by the same HCLWO algorithm. This process helps to prevent the noise and distortions present in the model. Finally, a numerical analysis is performed to investigate the effectiveness of the implemented process by contrasting numerous traditional models and algorithms.</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

An Adaptive Deep Learning-Based Channel Estimation with Hybrid Optimization-Aided Beamforming in mmwave Massive MIMO System

  • 期刊论文

  • 发表日期：

    2026年3月14日

查看全文
  • L. Rajeshwar Reddy,
  • B. Rajendra Naik

摘要

The fifth Generation (5G) communication device demands large connectivity with low delay and high data rates. One of the innovations to encounter these demands is millimeter Wave (mmwave) massive Multi-Input Multi-Output (MIMO). The hybrid beamforming and estimation of channel are taken for the multi-user mmwave MIMO device. In the modern days, various strategies have been recommended to model the hybrid beamforming and estimation of channels in mmwave MIMO devices. The conventional approaches concentrated on narrow-band channels. But, to efficiently employ the mmwave MIMO frameworks with higher bandwidth, there are modern tasks toward implementing the broadband hybrid beamforming mechanisms. The primary complexity in the hybrid beamforming is modeling a general beamformer, which is connected over the entire subcarriers while the weights of the digital beamformer are required to correspond to the subcarrier. In addition, the estimation of channels is complex because of the restricted Radio Frequency (RF) chains at the transceivers. Though numerous compressing sensing mechanisms have been implemented to rectify this issue by adopting sparsity frameworks, inherent channel sparsity, and practical works including beam squint, and power leakage can still create the original channel attributes that vary from desired techniques and lead to degradation in the functionality. Thus, in this work, a deep learning-assisted channel estimation technique with hybrid beamforming is implemented in the mmwave massive MIMO. At first, the necessary data is created for different scenarios to perform the channel estimation and hybrid beamforming process. Further, the Adaptive Convolutional Autoencoder (ACA) technique is supported to determine the channels in the mmwave massive MIMO device. Here, to increase the network efficiency and also provide better data transmission, the recommended ACA technique parameters are optimally tuned by the novel Hybrid Clouded Leopard and Walrus Optimization (HCLWO). Subsequently, the channel estimation and beam forming vectors are optimized by the same HCLWO algorithm. This process helps to prevent the noise and distortions present in the model. Finally, a numerical analysis is performed to investigate the effectiveness of the implemented process by contrasting numerous traditional models and algorithms.