Deep Learning for Uplink Spectral Efficiency in Cell-Free Massive MIMO Systems

Abstract

In this paper, we introduce a Deep Neural Network (DNN) to maximize the Proportional Fairness (PF) of the Spectral Efficiency (SE) of uplinks in Cell-Free (CF) massive Multiple-Input Multiple-Output (MIMO) systems. The problem of maximizing the PF of the SE is a nonconvex optimization problem in the design variables. We will develop a DNN which takes pilot sequences and largescale fading coefficients of the users as inputs and produces the outputs of optimal transmit powers. By consisting of densely residual connections between layers, the proposed DNN can efficiently exploit the hierarchical features of the input and motivates the feed-forward nature of DNN architecture. Experimental results showed that, compared to the conventional iterative optimization algorithm, the proposed DNN has excessively lower computational complexity with the trade-off approximately only 1% loss in the sum-rate and the fairness performance. This demonstrated that our proposed DNN is reasonably suitable for real-time signal processing in CF massive MIMO systems.Index Terms deep neural networks, proportional fairness, spectral efficiency, cell-free massive MIMO I. INTRODUCTIONIn recent years, Deep Neural Networks (DNNs) have obtained exceptional achievements in a variety of applications, especially in image processing. In [1], the authors proposed a deep network architecture named as Residual Network (ResNet) which consists of shortcut connections between layers to counteract the degradation problem in training significant DNNs. Numerical simulations demonstrated that in image classification and recognition, ResNets can achieve excellent results with significant depth, e.g. 1000 layers. In [2], the authors introduced a dense convolutional network, namely DenseNet. By feeding feature maps of previous layers to all subsequent layers, the DenseNets not only mitigate the gradient vanishing but also excessively deduct the trainable parameters. Numerical results shown that, compared to other state-of-the-art methods in the task of image classification, especially the ResNets [1], the DenseNets can obtained lower error rates with less computational efforts. In image restoration, deep

A Review of Energy Efficiency and Power Control Schemes in Ultra-Dense Cell-Free Massive MIMO Systems for Sustainable 6G Wireless Communication

“…Although groundbreaking advancements led by insightful ML solutions have been realized [205], the research in this line of interest is still in its infancy and requires further effort. Exploring yet undiscovered ML architectures that prioritize computational efficiency and energy consumption towards improved operating conditions may be interesting.…”