GPU-Based Implementation of Pruned Artificial Neural Networks for Digital Predistortion Linearization of Wideband Power Amplifiers

This paper presents a feature selection technique based on <inline-formula><tex-math notation="LaTeX">$\ell _{1}$</tex-math></inline-formula> regularization to select the most relevant weights of artificial neural networks (ANNs) for digital predistortion (DPD) line...

Full description

Saved in:
Bibliographic Details
Main Authors: Wantao Li, Raul Criado, William Thompson, Gabriel Montoro, Kevin Chuang, Pere L. Gilabert
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Microwaves
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10994208/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850144673094434816
author Wantao Li
Raul Criado
William Thompson
Gabriel Montoro
Kevin Chuang
Pere L. Gilabert
author_facet Wantao Li
Raul Criado
William Thompson
Gabriel Montoro
Kevin Chuang
Pere L. Gilabert
author_sort Wantao Li
collection DOAJ
description This paper presents a feature selection technique based on <inline-formula><tex-math notation="LaTeX">$\ell _{1}$</tex-math></inline-formula> regularization to select the most relevant weights of artificial neural networks (ANNs) for digital predistortion (DPD) linearization of wideband radio-frequency (RF) power amplifiers (PAs). The proposed pruning method is applied to the first hidden layer of a feed-forward real-valued time-delay neural network, commonly used for DPD purposes. In addition, this paper presents the ANN-based DPD implementation using a graphic processing unit (GPU) with compute unified device architecture (CUDA) units. Thanks to the proposed pruning strategy, it is possible to reduce the ANN complexity significantly, thereby achieving a higher data throughput with the GPU-based implementation. The trade-off among RF performance metrics, number of model parameters and throughput of the GPU implementation is evaluated considering the linearization of a high-efficiency pseudo-Doherty load modulated balanced amplifier (LMBA). The linearized PA operating at an RF frequency of 2 GHz delivers a mean output power of 40 dBm with approximately 50&#x0025; power efficiency when excited with 5G new radio (NR) signals with up to 200 MHz bandwidth and an 8 dB peak-to-average power ratio (PAPR). The real-time GPU implementation of the ANN-based DPD can meet the linearity specifications with a throughput circa 1 GSa/s.
format Article
id doaj-art-fe5faae54f8e45ae8992a409a53fd607
institution OA Journals
issn 2692-8388
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Microwaves
spelling doaj-art-fe5faae54f8e45ae8992a409a53fd6072025-08-20T02:28:18ZengIEEEIEEE Journal of Microwaves2692-83882025-01-015372673810.1109/JMW.2025.356042010994208GPU-Based Implementation of Pruned Artificial Neural Networks for Digital Predistortion Linearization of Wideband Power AmplifiersWantao Li0https://orcid.org/0000-0002-2634-6742Raul Criado1https://orcid.org/0009-0006-1064-520XWilliam Thompson2Gabriel Montoro3https://orcid.org/0000-0002-1328-4175Kevin Chuang4https://orcid.org/0000-0003-0805-5519Pere L. Gilabert5https://orcid.org/0000-0001-6183-6977Department of Signal Theory and Communications, University of Polit&#x00E8;cnica de Catalunya (UPC) - Barcelona Tech, Castelldefels, SpainDepartment of Signal Theory and Communications, University of Polit&#x00E8;cnica de Catalunya (UPC) - Barcelona Tech, Castelldefels, SpainAerospace, Defense and Communications Business Unit, Analog Devices Inc. (ADI), Wilmington, MA, USADepartment of Signal Theory and Communications, University of Polit&#x00E8;cnica de Catalunya (UPC) - Barcelona Tech, Castelldefels, SpainAerospace, Defense and Communications Business Unit, Analog Devices Inc. (ADI), Wilmington, MA, USADepartment of Signal Theory and Communications, University of Polit&#x00E8;cnica de Catalunya (UPC) - Barcelona Tech, Castelldefels, SpainThis paper presents a feature selection technique based on <inline-formula><tex-math notation="LaTeX">$\ell _{1}$</tex-math></inline-formula> regularization to select the most relevant weights of artificial neural networks (ANNs) for digital predistortion (DPD) linearization of wideband radio-frequency (RF) power amplifiers (PAs). The proposed pruning method is applied to the first hidden layer of a feed-forward real-valued time-delay neural network, commonly used for DPD purposes. In addition, this paper presents the ANN-based DPD implementation using a graphic processing unit (GPU) with compute unified device architecture (CUDA) units. Thanks to the proposed pruning strategy, it is possible to reduce the ANN complexity significantly, thereby achieving a higher data throughput with the GPU-based implementation. The trade-off among RF performance metrics, number of model parameters and throughput of the GPU implementation is evaluated considering the linearization of a high-efficiency pseudo-Doherty load modulated balanced amplifier (LMBA). The linearized PA operating at an RF frequency of 2 GHz delivers a mean output power of 40 dBm with approximately 50&#x0025; power efficiency when excited with 5G new radio (NR) signals with up to 200 MHz bandwidth and an 8 dB peak-to-average power ratio (PAPR). The real-time GPU implementation of the ANN-based DPD can meet the linearity specifications with a throughput circa 1 GSa/s.https://ieeexplore.ieee.org/document/10994208/Artificial neural networkdigital predistortiongraphics processing unitload-modulated balanced amplifiermodel-order reduction
spellingShingle Wantao Li
Raul Criado
William Thompson
Gabriel Montoro
Kevin Chuang
Pere L. Gilabert
GPU-Based Implementation of Pruned Artificial Neural Networks for Digital Predistortion Linearization of Wideband Power Amplifiers
IEEE Journal of Microwaves
Artificial neural network
digital predistortion
graphics processing unit
load-modulated balanced amplifier
model-order reduction
title GPU-Based Implementation of Pruned Artificial Neural Networks for Digital Predistortion Linearization of Wideband Power Amplifiers
title_full GPU-Based Implementation of Pruned Artificial Neural Networks for Digital Predistortion Linearization of Wideband Power Amplifiers
title_fullStr GPU-Based Implementation of Pruned Artificial Neural Networks for Digital Predistortion Linearization of Wideband Power Amplifiers
title_full_unstemmed GPU-Based Implementation of Pruned Artificial Neural Networks for Digital Predistortion Linearization of Wideband Power Amplifiers
title_short GPU-Based Implementation of Pruned Artificial Neural Networks for Digital Predistortion Linearization of Wideband Power Amplifiers
title_sort gpu based implementation of pruned artificial neural networks for digital predistortion linearization of wideband power amplifiers
topic Artificial neural network
digital predistortion
graphics processing unit
load-modulated balanced amplifier
model-order reduction
url https://ieeexplore.ieee.org/document/10994208/
work_keys_str_mv AT wantaoli gpubasedimplementationofprunedartificialneuralnetworksfordigitalpredistortionlinearizationofwidebandpoweramplifiers
AT raulcriado gpubasedimplementationofprunedartificialneuralnetworksfordigitalpredistortionlinearizationofwidebandpoweramplifiers
AT williamthompson gpubasedimplementationofprunedartificialneuralnetworksfordigitalpredistortionlinearizationofwidebandpoweramplifiers
AT gabrielmontoro gpubasedimplementationofprunedartificialneuralnetworksfordigitalpredistortionlinearizationofwidebandpoweramplifiers
AT kevinchuang gpubasedimplementationofprunedartificialneuralnetworksfordigitalpredistortionlinearizationofwidebandpoweramplifiers
AT perelgilabert gpubasedimplementationofprunedartificialneuralnetworksfordigitalpredistortionlinearizationofwidebandpoweramplifiers