Study of Robust Multiuser Scheduling and Power Allocation in Cell-Free MIMO Networks

TL;DR

Problem: robust downlink resource allocation in cell-free MIMO under imperfect CSI. Approach: a robust framework combining RC-ESG robust user scheduling and RGDPA robust gradient-descent power allocation with a linear MMSE precoder P, incorporating CSI errors into optimization via worst-case scheduling and gradient updates of power coefficients d, including a scaling step to meet the power constraint. Key contributions: RC-ESG algorithm for worst-case scheduling, RGDPA algorithm for robust power allocation with conditional MSE objective, and comprehensive simulations showing robustness gains under imperfect CSI. Impact: improved reliability and throughput in practical CF-mMIMO deployments where channel state information is imperfect.

Abstract

This paper introduces a robust resource allocation framework for the downlink of cell-free massive multi-input multi-output (CF-mMIMO) networks to address the effects caused by imperfect channel state information (CSI). In particular, the proposed robust resource allocation framework includes a robust user scheduling algorithm to optimize the network's sum-rate and a robust power allocation technique aimed at minimizing the mean square error (MSE) for a network with a linear precoder. Unlike non-robust resource allocation techniques, the proposed robust strategies effectively counteract the effects of imperfect CSI, enhancing network efficiency and reliability. Simulation results show a significant improvement in network performance obtained by the proposed approaches, highlighting the impact of robust resource allocation in wireless networks.

Figures (2)

• Figure 1: Comparison of the C-ESG and RC-ESG in PCSI and ICSI CF networks when EPL is used and $\alpha=0.15$ for ICSI case, $M=64$, $K=32$ and $n=16$.
• Figure 2: Comparison of the RGDPA and GDPA power allocation techniques when C-ESG user scheduling is used, $\alpha=0.15$ for ICSI case, $M=64$, $K=32$ and $n=16$.