Graph neural networks improve antenna polarization for reliable wireless links
GNN-Based Polarforming for Multi-User MISO Short-Packet URLLC under Imperfect CSI
Information Theory
Summary
Wireless communication often needs to send data quickly and without errors, but signals can be disrupted by imperfect knowledge of the wireless channel and how antennas are oriented. The authors study a system where antennas can change their polarization (the orientation of the signal waves) to better match the environment. They use a specially designed graph neural network to decide how to set antenna parameters and beam signals to multiple users at once, even when channel information is uncertain. Their approach improves the amount of data reliably transmitted and reduces errors compared to traditional fixed antenna orientations.
What this means in practice
- •For wireless network engineers: Enhance multi-user wireless base stations with real-time adaptive antenna polarization and beamforming to improve data rates and reliability under uncertain channel conditions.
- •For industrial automation teams: Support ultra-reliable low-latency device communication in factory environments by dynamically adjusting antenna properties to maintain link quality despite interference and channel variations.
Authors
Zahra Mehrzad, Hamed Aghaei-Karkaj, Rahman Saadat Yeganeh, Kamran Ebrahimi, Mohammad Robat Mili, Symeon Chatzinotas, Ioannis Krikidis
Abstract
This paper investigates polarization-aware transmission for multi-user multiple-input single-output (MU-MISO) short-packet ultra-reliable low-latency communications (URLLC) under imperfect channel state information (CSI). We consider a system in which the base station (BS) and users are equipped with polarization-reconfigurable antennas that enable adaptive polarization states through controllable polarization coefficients. A multi-objective optimization problem is formulated to jointly maximize the finite-blocklength (FBL) achievable sum rate and minimize the maximum decoding error probability (DEP), subject to transmit-power, latency, reliability, and discrete polarization-control constraints. The resulting multi-objective problem is scalarized using a normalized weighted-sum utility. To enable low-complexity online decision-making, a heterogeneous graph neural network (GNN) is developed to learn the joint mapping from estimated polarized CSI to digital beamforming and transmit/receive polarforming vectors (PFVs) while accounting for the system constraints. Numerical results demonstrate that the proposed GNN-based polarforming (PF) framework substantially improves the FBL sum rate while reducing the maximum DEP compared with conventional fixed-polarization schemes, particularly under channel depolarization and imperfect CSI. This highlights the potential of adaptive polarization control for reliable low-latency transmission.