EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Sparsely Gated Mixture of Experts Neural Network For Linearization of RF Power Amplifiers

Arne Fischer-Bühner, Alberto Brihuega, Lauri Anttila, Matias Turunen, Vishnu Unnikrishnan, Manil Dev Gomony, Mikko Valkama

IEEE Transactions on Microwave Theory and Techniques · 2024-07 · 72 (7) · 4367-4382

DOI: 10.1109/tmtt.2023.3341616 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Microwave Theory and Techniques, 2024-07. Verfasst von Arne Fischer-Bühner und 6 weiteren Autor:innen. Die Arbeit wurde bisher 22-mal zitiert (Crossref) und nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: European Union’s Horizon 2020 Research and Innovation Program through the Marie Skłodowska-Curie, Academy of Finland.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This article presents a piecewise neural network (NN) with dynamic sparsity for modeling and linearization of radio frequency (RF) power amplifiers (PAs). A mixture of experts NN (MENN) approach is employed to combine several smaller real-valued time-delay NNs (RVTDNNs) by means of a gating NN. Furthermore, we complement the MENN framework with top- $K$ sparse gating, such that only a subset of experts is activated during each sample inference, reducing the computational complexity at run time. An end-to-end training approach is presented, to optimize the gating alongside with specializing the expert NNs, enabling the experts to collaborate. We experimentally investigate the scaleability of the proposed model in terms of modeling accuracy and linearization performance, as well as run time and model complexity, using RF measurements with two different gallium-nitride Doherty PAs at 1.8 and 3.5 GHz, respectively. Our experiments confirm a significant reduction in run-time complexity due to the sparse gating, with only a small penalty on accuracy, linearization capability and scaleability. Furthermore, the proposed approach is shown to offer favorable complexity-performance trade-offs, outperforming the existing state-of-the-art.

Förderung

  • European Union’s Horizon 2020 Research and Innovation Program through the Marie Skłodowska-Curie
  • Academy of Finland

Zitieren

APA

Fischer-Bühner, A., Brihuega, A., Anttila, L., Turunen, M., Unnikrishnan, V., Gomony, M. D., & Valkama, M. (2024). Sparsely Gated Mixture of Experts Neural Network For Linearization of RF Power Amplifiers. IEEE Transactions on Microwave Theory and Techniques, 72(7), 4367-4382. https://doi.org/10.1109/tmtt.2023.3341616

Vancouver

Fischer-Bühner A, Brihuega A, Anttila L, Turunen M, Unnikrishnan V, Gomony MD, et al. Sparsely Gated Mixture of Experts Neural Network For Linearization of RF Power Amplifiers. IEEE Transactions on Microwave Theory and Techniques. 2024;72(7):4367-4382. doi:10.1109/tmtt.2023.3341616

BibTeX
@article{FischerBhner2024,
  author = {Fischer-Bühner, Arne and Brihuega, Alberto and Anttila, Lauri and Turunen, Matias and Unnikrishnan, Vishnu and Gomony, Manil Dev and Valkama, Mikko},
  title = {Sparsely Gated Mixture of Experts Neural Network For Linearization of RF Power Amplifiers},
  journal = {IEEE Transactions on Microwave Theory and Techniques},
  year = {2024},
  volume = {72},
  number = {7},
  pages = {4367--4382},
  doi = {10.1109/tmtt.2023.3341616}
}
Kostenlos starten →