2026-04-18T05:31:08Zhttps://recercat.cat/oai/request

oai:recercat.cat:2117/3703162026-01-21T09:52:01Zcom_2072_1033col_2072_452950

urn:hdl:2117/370316 Agnostic envelope linearization of dynamically supplied power amplifiers for mobile terminals Li, Wantao Montoro López, Gabriel Gilabert Pinal, Pere Lluís Àrees temàtiques de la UPC::Enginyeria de la telecomunicació Power amplifiers Digital predistortion Envelope tracking Power amplifier RF leakage Amplificadors de potència This paper presents an envelope linearization technique to compensate for the nonlinear distortion of envelope tracking (ET) power amplifiers (PAs) for 5G new radio (NR) mobile terminals. The proposed envelope optimization (EOPT) method is agnostic of the nonlinear distortion generated in the envelope supply path and can compensate for the nonlinear distortion at the ET PA output without the need to monitor the output at the envelope tracking modulator (ETM). The linearization system in the envelope path is based on the envelope generalized memory polynomial (EGMP) behavioral model. Since the ETM output is not available, an iterative nonlinear least squares solution inspired in the deep deterministic policy gradient (DDPG) algorithm is proposed to extract the coefficients of the EGMP model. The EOPT method is validated on a system-on-chip (SoC) ET PA board designed for mobile terminal applications. Experimental results show the suitability of the proposed method to guarantee the linearity requirements (i.e., adjacent channel power ratio below -36 dBc) with 16.8% of power efficiency when operating the ET PA with 5G new radio test signals of 60 MHz bandwidth operating at 2.55 GHz (band 7). The linearization performance of the proposed EOPT method is comparable to the envelope leakage cancellation (ELC) approach (but saving the need for an analog to digital converter to monitor the ETM output), and can outperform a conventional I-Q digital predistorter based on the generalized memory polynomial (GMP) behavioral model. This research was funded by Huawei Technologies from July 2020 to August 2021; and supported in part by the project PID2020-113832RB-C21 funded by MCIN/AEI/10.13039/50110001103 and in part by the Government of Catalonia and the European Social Fund under Grant 2021-FI-B-137. Peer Reviewed Postprint (published version) 2022-05-16 Article https://www.mdpi.com/1424-8220/22/10/3773 info:eu-repo/grantAgreement/AEI/Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020/PID2020-113832RB-C21/ES/DISPOSITIVOS ASISTIDOS POR TECNICAS DE DEEP Y MACHINE LEARNING PARA TRANSCEPTORES DE RADIOFRECUENCIA ALTAMENTE EFICIENTES/ http://creativecommons.org/licenses/by-nc-nd/4.0/ Open Access Attribution-NonCommercial-NoDerivatives 4.0 International Multidisciplinary Digital Publishing Institute (MDPI)