Learning-Induced Channel Extrapolation for Fluid Antenna Systems Using Asymmetric Graph Masked Autoencoder

Advanced search
Browse by:

Department | Year

UCL Theses | Latest

Deposit your research

Learning-Induced Channel Extrapolation for Fluid Antenna Systems Using Asymmetric Graph Masked Autoencoder

Zhang, H; Wang, J; Wang, C; Wang, CC; Wong, KK; Wang, B; Chae, CB; (2024) Learning-Induced Channel Extrapolation for Fluid Antenna Systems Using Asymmetric Graph Masked Autoencoder. IEEE Wireless Communications Letters 10.1109/LWC.2024.3386153. (In press). Green open access

Preview

PDF
WCL2024-0282-R1.pdf - Accepted Version
Download (12MB) | Preview

Abstract

The emerging fluid antenna systems (FASs) enable position-flexible antennas that provide a new degree of freedom for more diversity and multiplexing benefits. However, the challenge for high-resolution FAS is the acquisition of complete and accurate channel state information (CSI) within a coherence time period. In this correspondence, we study deep learning methods for CSI extrapolation in high-resolution FAS with less complexity and greater generalization ability. In so doing, we then contrive a customized solution, referred to as an asymmetric graph masked autoencoder (AGMAE), specifically designed for spatial channel extrapolation in FAS. This technique incorporates an attention mechanism, an asymmetric masked autoencoder architecture to reduce computational complexity, and utilizes the local diffusion mechanism of graph neural networks to enhance generalization. Simulation results validate the effectiveness and generality of the proposed method for the CSI acquisition of FAS.

Type:	Article
Title:	Learning-Induced Channel Extrapolation for Fluid Antenna Systems Using Asymmetric Graph Masked Autoencoder
Open access status:	An open access version is available from UCL Discovery
DOI:	10.1109/LWC.2024.3386153
Publisher version:	http://dx.doi.org/10.1109/lwc.2024.3386153
Language:	English
Additional information:	This version is the author accepted manuscript. For information on re-use, please refer to the publisher's terms and conditions.
UCL classification:	UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Electronic and Electrical Eng
URI:	https://discovery.ucl.ac.uk/id/eprint/10191114