Comparison of a genetic programming approach with ANFIS for power amplifier behavioral modeling and FPGA implementation

José Alejandro Galaviz-Aguilar, Patrick Roblin, José Ricardo Cárdenas-Valdez, Emigdio Z-Flores, Leonardo Trujillo, José Cruz Nuñez-Pérez, Oliver Schütze

Research output: Contribution to journalArticleResearchpeer-review

1 Citation (Scopus)

Abstract

© 2017, Springer-Verlag GmbH Germany, part of Springer Nature. Accurate modeling of power amplifiers (PA) is of upmost importance in the design process of wireless communication systems where a high linearity and efficiency is required. To deal with the nonlinear behavior of PAs effectively a linearization stage is applied to minimize the distortions of in-band and adjacent transmission channels, which translate to an improvement of the signal integrity and the operation cost of the transmitter system. This paper presents a method based on genetic programming with a local search heuristic (GP-LS) to emulate the electrical memory effects by using the characteristic conversion curves of the radio frequency (RF) PA NXP Semiconductor of 10 W GaN HEMT working at 2.34 GHz. This method is compared with an Adaptive Neuro-Fuzzy Inference Systems (ANFIS) through several performance metrics (NMSE, MAE and correlation coefficient), with GP-LS achieving a better modeling accuracy. Moreover, the models produced by GP-LS permit a reduction in the required hardware resources, when it is implemented on a Field-Programmable Gate Array through the DSP Builder tool. The models are derived using a data-driven approach, posed in two different ways. Firstly, experiments are performed using a testbed Arria V GX for a flexible vector signal generation that provides the raw data of the PA characterization using an LTE-Advanced signal with 10-MHz bandwidth. Secondly, the modeling is derived from a filtered version of the data and then adding a high-frequency signal as a post processing step to approximate the true behavior of the system. In both cases, the models are generated with ANFIS and GP-LS, performing extensive logic-based simulations and implementing the models on a Cyclone III development board. Both approaches are compared based on accuracy and required hardware resources, with GP-LS substantially outperforming ANFIS. These results suggest that the GP-LS models can be implemented in a digital predistortion chain and used in the linearization stage for a RF-PA.
Original languageAmerican English
Pages (from-to)2463-2481
Number of pages2214
JournalSoft Computing
DOIs
StatePublished - 15 Apr 2019

Fingerprint

Behavioral Modeling
FPGA Implementation
Adaptive Neuro-fuzzy Inference System
Power Amplifier
Genetic programming
Fuzzy inference
Genetic Programming
Power amplifiers
Field programmable gate arrays (FPGA)
Radio frequency amplifiers
Linearization
Modeling
Hardware
LTE-advanced
Resources
Model
Memory Effect
High electron mobility transistors
Performance Metrics
Testbeds

Cite this

Galaviz-Aguilar, J. A., Roblin, P., Cárdenas-Valdez, J. R., Z-Flores, E., Trujillo, L., Nuñez-Pérez, J. C., & Schütze, O. (2019). Comparison of a genetic programming approach with ANFIS for power amplifier behavioral modeling and FPGA implementation. Soft Computing, 2463-2481. https://doi.org/10.1007/s00500-017-2941-8
Galaviz-Aguilar, José Alejandro ; Roblin, Patrick ; Cárdenas-Valdez, José Ricardo ; Z-Flores, Emigdio ; Trujillo, Leonardo ; Nuñez-Pérez, José Cruz ; Schütze, Oliver. / Comparison of a genetic programming approach with ANFIS for power amplifier behavioral modeling and FPGA implementation. In: Soft Computing. 2019 ; pp. 2463-2481.
@article{74ccd1436a7245b287159ce5f2b9d991,
title = "Comparison of a genetic programming approach with ANFIS for power amplifier behavioral modeling and FPGA implementation",
abstract = "{\circledC} 2017, Springer-Verlag GmbH Germany, part of Springer Nature. Accurate modeling of power amplifiers (PA) is of upmost importance in the design process of wireless communication systems where a high linearity and efficiency is required. To deal with the nonlinear behavior of PAs effectively a linearization stage is applied to minimize the distortions of in-band and adjacent transmission channels, which translate to an improvement of the signal integrity and the operation cost of the transmitter system. This paper presents a method based on genetic programming with a local search heuristic (GP-LS) to emulate the electrical memory effects by using the characteristic conversion curves of the radio frequency (RF) PA NXP Semiconductor of 10 W GaN HEMT working at 2.34 GHz. This method is compared with an Adaptive Neuro-Fuzzy Inference Systems (ANFIS) through several performance metrics (NMSE, MAE and correlation coefficient), with GP-LS achieving a better modeling accuracy. Moreover, the models produced by GP-LS permit a reduction in the required hardware resources, when it is implemented on a Field-Programmable Gate Array through the DSP Builder tool. The models are derived using a data-driven approach, posed in two different ways. Firstly, experiments are performed using a testbed Arria V GX for a flexible vector signal generation that provides the raw data of the PA characterization using an LTE-Advanced signal with 10-MHz bandwidth. Secondly, the modeling is derived from a filtered version of the data and then adding a high-frequency signal as a post processing step to approximate the true behavior of the system. In both cases, the models are generated with ANFIS and GP-LS, performing extensive logic-based simulations and implementing the models on a Cyclone III development board. Both approaches are compared based on accuracy and required hardware resources, with GP-LS substantially outperforming ANFIS. These results suggest that the GP-LS models can be implemented in a digital predistortion chain and used in the linearization stage for a RF-PA.",
author = "Galaviz-Aguilar, {Jos{\'e} Alejandro} and Patrick Roblin and C{\'a}rdenas-Valdez, {Jos{\'e} Ricardo} and Emigdio Z-Flores and Leonardo Trujillo and Nu{\~n}ez-P{\'e}rez, {Jos{\'e} Cruz} and Oliver Sch{\"u}tze",
year = "2019",
month = "4",
day = "15",
doi = "10.1007/s00500-017-2941-8",
language = "American English",
pages = "2463--2481",
journal = "Soft Computing",
issn = "1432-7643",
publisher = "Springer Verlag",

}

Galaviz-Aguilar, JA, Roblin, P, Cárdenas-Valdez, JR, Z-Flores, E, Trujillo, L, Nuñez-Pérez, JC & Schütze, O 2019, 'Comparison of a genetic programming approach with ANFIS for power amplifier behavioral modeling and FPGA implementation', Soft Computing, pp. 2463-2481. https://doi.org/10.1007/s00500-017-2941-8

Comparison of a genetic programming approach with ANFIS for power amplifier behavioral modeling and FPGA implementation. / Galaviz-Aguilar, José Alejandro; Roblin, Patrick; Cárdenas-Valdez, José Ricardo; Z-Flores, Emigdio; Trujillo, Leonardo; Nuñez-Pérez, José Cruz; Schütze, Oliver.

In: Soft Computing, 15.04.2019, p. 2463-2481.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Comparison of a genetic programming approach with ANFIS for power amplifier behavioral modeling and FPGA implementation

AU - Galaviz-Aguilar, José Alejandro

AU - Roblin, Patrick

AU - Cárdenas-Valdez, José Ricardo

AU - Z-Flores, Emigdio

AU - Trujillo, Leonardo

AU - Nuñez-Pérez, José Cruz

AU - Schütze, Oliver

PY - 2019/4/15

Y1 - 2019/4/15

N2 - © 2017, Springer-Verlag GmbH Germany, part of Springer Nature. Accurate modeling of power amplifiers (PA) is of upmost importance in the design process of wireless communication systems where a high linearity and efficiency is required. To deal with the nonlinear behavior of PAs effectively a linearization stage is applied to minimize the distortions of in-band and adjacent transmission channels, which translate to an improvement of the signal integrity and the operation cost of the transmitter system. This paper presents a method based on genetic programming with a local search heuristic (GP-LS) to emulate the electrical memory effects by using the characteristic conversion curves of the radio frequency (RF) PA NXP Semiconductor of 10 W GaN HEMT working at 2.34 GHz. This method is compared with an Adaptive Neuro-Fuzzy Inference Systems (ANFIS) through several performance metrics (NMSE, MAE and correlation coefficient), with GP-LS achieving a better modeling accuracy. Moreover, the models produced by GP-LS permit a reduction in the required hardware resources, when it is implemented on a Field-Programmable Gate Array through the DSP Builder tool. The models are derived using a data-driven approach, posed in two different ways. Firstly, experiments are performed using a testbed Arria V GX for a flexible vector signal generation that provides the raw data of the PA characterization using an LTE-Advanced signal with 10-MHz bandwidth. Secondly, the modeling is derived from a filtered version of the data and then adding a high-frequency signal as a post processing step to approximate the true behavior of the system. In both cases, the models are generated with ANFIS and GP-LS, performing extensive logic-based simulations and implementing the models on a Cyclone III development board. Both approaches are compared based on accuracy and required hardware resources, with GP-LS substantially outperforming ANFIS. These results suggest that the GP-LS models can be implemented in a digital predistortion chain and used in the linearization stage for a RF-PA.

AB - © 2017, Springer-Verlag GmbH Germany, part of Springer Nature. Accurate modeling of power amplifiers (PA) is of upmost importance in the design process of wireless communication systems where a high linearity and efficiency is required. To deal with the nonlinear behavior of PAs effectively a linearization stage is applied to minimize the distortions of in-band and adjacent transmission channels, which translate to an improvement of the signal integrity and the operation cost of the transmitter system. This paper presents a method based on genetic programming with a local search heuristic (GP-LS) to emulate the electrical memory effects by using the characteristic conversion curves of the radio frequency (RF) PA NXP Semiconductor of 10 W GaN HEMT working at 2.34 GHz. This method is compared with an Adaptive Neuro-Fuzzy Inference Systems (ANFIS) through several performance metrics (NMSE, MAE and correlation coefficient), with GP-LS achieving a better modeling accuracy. Moreover, the models produced by GP-LS permit a reduction in the required hardware resources, when it is implemented on a Field-Programmable Gate Array through the DSP Builder tool. The models are derived using a data-driven approach, posed in two different ways. Firstly, experiments are performed using a testbed Arria V GX for a flexible vector signal generation that provides the raw data of the PA characterization using an LTE-Advanced signal with 10-MHz bandwidth. Secondly, the modeling is derived from a filtered version of the data and then adding a high-frequency signal as a post processing step to approximate the true behavior of the system. In both cases, the models are generated with ANFIS and GP-LS, performing extensive logic-based simulations and implementing the models on a Cyclone III development board. Both approaches are compared based on accuracy and required hardware resources, with GP-LS substantially outperforming ANFIS. These results suggest that the GP-LS models can be implemented in a digital predistortion chain and used in the linearization stage for a RF-PA.

UR - https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85034854430&origin=inward

UR - https://www.scopus.com/inward/citedby.uri?partnerID=HzOxMe3b&scp=85034854430&origin=inward

U2 - 10.1007/s00500-017-2941-8

DO - 10.1007/s00500-017-2941-8

M3 - Article

SP - 2463

EP - 2481

JO - Soft Computing

JF - Soft Computing

SN - 1432-7643

ER -

Galaviz-Aguilar JA, Roblin P, Cárdenas-Valdez JR, Z-Flores E, Trujillo L, Nuñez-Pérez JC et al. Comparison of a genetic programming approach with ANFIS for power amplifier behavioral modeling and FPGA implementation. Soft Computing. 2019 Apr 15;2463-2481. https://doi.org/10.1007/s00500-017-2941-8