Online Adaptive Real-Time Beamforming Design for Dynamic Environments in Cell-Free Systems

TL;DR

This work tackles real-time beamforming in dynamic cell-free wireless systems where channel statistics and network size vary over time. It introduces HGNet, a CNN-based framework with a high-generalization beamforming module that extracts robust features and adapts to changing numbers of APs and users, along with a theoretical bound based on G-MMD to explain improved generalization. An online adaptive updating (OAU) procedure then updates a small fraction of BN parameters using an information-entropy objective to maintain millisecond-scale reflexes in live operation. Empirical results show HGNet achieving higher sum-rate with lower computation time than state-of-the-art baselines, and OAU providing further gains while maintaining real-time performance.

Abstract

In this paper, we consider real-time beamforming design for dynamic wireless environments with varying channels and different numbers of access points (APs) and users in cell-free systems. Specifically, a sum-rate maximization optimization problem is formulated for the beamforming design in dynamic wireless environments of cell-free systems. To efficiently solve it, a high-generalization network (HGNet) is proposed to adapt to the changing numbers of APs and users. Then, a high-generalization beamforming module is also designed in HGNet to extract the valuable features for the varying channels, and we theoretically prove that such a high-generalization beamforming module is able to reduce the upper bound of the generalization error. Subsequently, by online adaptively updating about 3% of the parameters of HGNet, an online adaptive updating (OAU) algorithm is proposed to enable the online adaptive real-time beamforming design for improving the sum rate. Numerical results demonstrate that the proposed HGNet with OAU algorithm achieves a higher sum rate with a lower computational cost on the order of milliseconds, thus realizing the real-time beamforming design for dynamic wireless environments in cell-free systems.

Figures (11)

• Figure 1: An illustration for dynamic wireless environments with the varying channels and the different numbers of APs and users.
• Figure 2: Proposed HGNet.
• Figure 3: Plot of training loss with the number of iterations.
• Figure 4: Comparative results of average generalization sum rate across three channels under $Q_t=24$, $I_t=24$.
• Figure 5: Comparative results of average computation time across three channels under $Q_t=24$, $I_t=24$.