On Feedback Sample Selection Methods Allowing Lightweight Digital Predistorter Adaptation

On Feedback Sample Selection Methods Allowing Lightweight Digital Predistorter Adaptation

WoS Article

In modern communication systems advanced techniques such as digital predistortion (DPD) are required to satisfy stringent demands on transmitter linearity and efficiency. DPD, however, increases the hardware and computational complexity of transmitters. In this article we show that the computational complexity of DPD adaptation can be drastically reduced if a low number of samples is properly selected for DPD adaptation. For this purpose we propose methods of sample selection for DPD adaptation. Among the proposed methods, the highest computational complexity reduction is achieved by a method based on the histogram of signal magnitudes, optimised with respect to characteristics of the radio frequency power amplifier and of the transmitted signal. Simulations indicate that this proposed method can reduce the computational complexity of DPD adaptation by a factor of up to 400 while the linearisation performance of conventional methods is preserved. Besides simulations for three models of distinct power amplifiers, measurements on a real power amplifier further verify the linearisation capabilities of the proposed methods.

In modern communication systems advanced techniques such as digital predistortion (DPD) are required to satisfy stringent demands on transmitter linearity and efficiency. DPD, however, increases the hardware and computational complexity of transmitters. In this article we show that the computational complexity of DPD adaptation can be drastically reduced if a low number of samples is properly selected for DPD adaptation. For this purpose we propose methods of sample selection for DPD adaptation. Among the proposed methods, the highest computational complexity reduction is achieved by a method based on the histogram of signal magnitudes, optimised with respect to characteristics of the radio frequency power amplifier and of the transmitted signal. Simulations indicate that this proposed method can reduce the computational complexity of DPD adaptation by a factor of up to 400 while the linearisation performance of conventional methods is preserved. Besides simulations for three models of distinct power amplifiers, measurements on a real power amplifier further verify the linearisation capabilities of the proposed methods.

power amplifier linearisation, digital predistortion, low complexity adaptation, histogram-based sample selection, sample subset reduction, direct learning architecture

power amplifier linearisation, digital predistortion, low complexity adaptation, histogram-based sample selection, sample subset reduction, direct learning architecture

KRÁL, J.; GÖTTHANS, T.; MARŠÁLEK, R.; HARVÁNEK, M.; RUPP, M.

2021

25.06.2020

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC

PISCATAWAY

1558-0806

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS

67

6

United States of America

1976

1988

13

https://ieeexplore.ieee.org/document/9018116

165453_UREL_Kral_merge_red