Learning-Based Precoding for Zak-OTFS
IEEE Communications Letters · 2026 · 30 · 2730-2734
DOI: 10.1109/lcomm.2026.3716739 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Communications Letters, 2026. Verfasst von Naveed Bin Nazir und 1 weiteren Autor:innen. Die Arbeit nennt 7 Literaturangaben. Gefördert von: J. C. Bose National Fellowship and the NSF-DST project (International Cooperation Division), Department of Science and Technology, Government of India.
Abstract (Originalsprache)
A recently proposed twisted convolution (TC) precoder for Zak-OTFS enables 1-tap equalization at the receiver by maximizing the per-carrier signal-to-interference-plus-noise ratio (SINR). This precoder has noticeable leakage around the target tap which limits performance. In this letter, we tackle this problem in two stages. First, at the transmitter, we propose two learning-based TC precoders, namely, a gradient-based perturbation learning (GBPL) precoder and a liquid neural network (LNN)-based precoder, which suppress the leakage better than the SINR-max precoder, with the LNN precoder producing the strongest localization. Second, at the receiver, unlike conventional receivers that separate channel estimation and symbol detection tasks, we design a lightweight convolutional neural network (CNN) that performs both tasks jointly by simultaneously denoising the received frame and mitigating the residual leakage. Simulation results for fractional delay-Doppler channels show that the proposed precoders outperform the SINR-max precoder, and the CNN receiver offers additional performance gains by efficiently handling remaining leakage.
Förderung
- J. C. Bose National Fellowship and the NSF-DST project (International Cooperation Division), Department of Science and Technology, Government of India
Zitieren
APA
Nazir, N. B., & Chockalingam, A. (2026). Learning-Based Precoding for Zak-OTFS. IEEE Communications Letters, 30, 2730-2734. https://doi.org/10.1109/lcomm.2026.3716739
Vancouver
Nazir NB, Chockalingam A. Learning-Based Precoding for Zak-OTFS. IEEE Communications Letters. 2026;30:2730-2734. doi:10.1109/lcomm.2026.3716739
BibTeX
@article{Nazir2026,
author = {Nazir, Naveed Bin and Chockalingam, Ananthanarayanan},
title = {Learning-Based Precoding for Zak-OTFS},
journal = {IEEE Communications Letters},
year = {2026},
volume = {30},
pages = {2730--2734},
doi = {10.1109/lcomm.2026.3716739}
}
EUVIMED