EUVIMEDEUVIMED entdecken
Crossref

Low‐Complexity Neural Network Architecture for High‐Accuracy Wideband Power Amplifier Modeling and Digital Predistortion Linearization

Haithem Rezgui, Amine Ben Salem, Rim Barrak, Adel Ghazel

International Journal of Circuit Theory and Applications · 28.04.2026

DOI: 10.1002/cta.70448 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Circuit Theory and Applications, online erschienen am 28.04.2026. Verfasst von Haithem Rezgui und 3 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In this paper, we introduce a new behavioral model for wideband RF power amplifier (PA) based on enhanced real‐valued time‐delay neural network (ERVTDNN) architecture. The core innovation is the introduction of leading terms into the input basis function, which generates crucial cross‐effects with memory terms. Unlike conventional models, this unique architecture enriches the basis function with both odd and even order terms, significantly improving prediction accuracy and linearization capabilities. The advantages of the proposed neural network architecture over existing models in the literature are demonstrated through fundamental analysis. The modeling and linearization performances of the ERVTDNN‐based architecture are validated using experimental PA measurements. The proposed topology is applied to nonlinear distortion compensation in a 25‐W gallium nitride (GaN) RF PA using digital predistortion (DPD). Our results compellingly demonstrate that the proposed model drastically reduces computational complexity by over 45% compared with relevant reference models, while achieving a 15‐dB reduction in out‐of‐band distortion, as measured by the adjacent channel power ratio (ACPR), relative to the no DPD configuration.

Zitieren

APA

Rezgui, H., Salem, A. B., Barrak, R., & Ghazel, A. (2026). Low‐Complexity Neural Network Architecture for High‐Accuracy Wideband Power Amplifier Modeling and Digital Predistortion Linearization. International Journal of Circuit Theory and Applications. https://doi.org/10.1002/cta.70448

Vancouver

Rezgui H, Salem AB, Barrak R, Ghazel A. Low‐Complexity Neural Network Architecture for High‐Accuracy Wideband Power Amplifier Modeling and Digital Predistortion Linearization. International Journal of Circuit Theory and Applications. 2026. doi:10.1002/cta.70448

BibTeX
@article{Rezgui2026,
  author = {Rezgui, Haithem and Salem, Amine Ben and Barrak, Rim and Ghazel, Adel},
  title = {Low‐Complexity Neural Network Architecture for High‐Accuracy Wideband Power Amplifier Modeling and Digital Predistortion Linearization},
  journal = {International Journal of Circuit Theory and Applications},
  year = {2026},
  doi = {10.1002/cta.70448}
}
Kostenlos starten →