Deep Unfolded Iterative Shrinkage-Thresholding With Mask-Guided Prior for OFDM Channel Estimation
IEEE Communications Letters · 2026 · 30 · 2635-2639
DOI: 10.1109/lcomm.2026.3713385 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Communications Letters, 2026. Verfasst von Mengyang Wu und 2 weiteren Autor:innen. Die Arbeit nennt 10 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Guangxi Natural Science Foundation, Project of Guangxi Wireless Broadband Communication and Signal Processing Key Laboratory.
Abstract (Originalsprache)
Deep learning models for channel estimation in Orthogonal Frequency Division Multiplexing (OFDM) systems often suffer from limited interpretability and weak generalization across varying pilot configurations and frequency-selective fading conditions. To address this issue, a model-driven deep unfolded Iterative Shrinkage-Thresholding Algorithm (ISTA) framework, termed MGP-ISTA, is proposed, in which a mask-guided partial-convolution prior is incorporated to alleviate interpolation errors, and a dual-path CNN is adopted as a trainable proximal operator. Simulation results show that the proposed method consistently outperforms the LS-Spline and ChannelNet baselines under different pilot configurations and SNR conditions, while remaining robust in frequency-selective fading environments.
Förderung
- National Natural Science Foundation of China
- Guangxi Natural Science Foundation
- Project of Guangxi Wireless Broadband Communication and Signal Processing Key Laboratory
Zitieren
APA
Wu, M., Zhou, H., & Sun, F. (2026). Deep Unfolded Iterative Shrinkage-Thresholding With Mask-Guided Prior for OFDM Channel Estimation. IEEE Communications Letters, 30, 2635-2639. https://doi.org/10.1109/lcomm.2026.3713385
Vancouver
Wu M, Zhou H, Sun F. Deep Unfolded Iterative Shrinkage-Thresholding With Mask-Guided Prior for OFDM Channel Estimation. IEEE Communications Letters. 2026;30:2635-2639. doi:10.1109/lcomm.2026.3713385
BibTeX
@article{Wu2026,
author = {Wu, Mengyang and Zhou, Hong and Sun, Fengyuan},
title = {Deep Unfolded Iterative Shrinkage-Thresholding With Mask-Guided Prior for OFDM Channel Estimation},
journal = {IEEE Communications Letters},
year = {2026},
volume = {30},
pages = {2635--2639},
doi = {10.1109/lcomm.2026.3713385}
}
EUVIMED