A Kronecker Product-Based Module for Digital Predistortion Neural Network for the Wide Bandwidth Power Amplifier
IEEE Microwave and Wireless Technology Letters · 2026-01 · 36 (1) · 143-146
DOI: 10.1109/lmwt.2025.3613530 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Microwave and Wireless Technology Letters, 2026-01. Verfasst von Jianfeng Shao und 2 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 12 Literaturangaben. Gefördert von: ZTE Corporation.
Abstract (Originalsprache)
The strong data fitting ability of neural networks (NNs) makes them widely used in research on digital predistortion (DPD) for mitigating the nonlinear effects of power amplifiers (PAs). However, existing NN needs a large number of coefficients to achieve satisfactory performance, especially in the wider bandwidth scenarios. To address the high complexity demand of conventional DPD NNs, a Kronecker product-based module is proposed. It is structured using linear layers and Kronecker product operations to compose the nonlinear terms based on the memory polynomial function. Verification using 100- and 400-MHz datasets demonstrates the effectiveness of the KP-based module. It enables conventional DPD NNs to achieve better performance while preventing it degradation under lower NN complexity. Afterward, the stabilization validation also demonstrates the KP-based module helps DPD NNs to achieve more stable performance.
Förderung
- ZTE Corporation
Zitieren
APA
Shao, J., Hong, X., & Wang, W. (2026). A Kronecker Product-Based Module for Digital Predistortion Neural Network for the Wide Bandwidth Power Amplifier. IEEE Microwave and Wireless Technology Letters, 36(1), 143-146. https://doi.org/10.1109/lmwt.2025.3613530
Vancouver
Shao J, Hong X, Wang W. A Kronecker Product-Based Module for Digital Predistortion Neural Network for the Wide Bandwidth Power Amplifier. IEEE Microwave and Wireless Technology Letters. 2026;36(1):143-146. doi:10.1109/lmwt.2025.3613530
BibTeX
@article{Shao2026,
author = {Shao, Jianfeng and Hong, Xi and Wang, Wenjie},
title = {A Kronecker Product-Based Module for Digital Predistortion Neural Network for the Wide Bandwidth Power Amplifier},
journal = {IEEE Microwave and Wireless Technology Letters},
year = {2026},
volume = {36},
number = {1},
pages = {143--146},
doi = {10.1109/lmwt.2025.3613530}
}
EUVIMED