An optimized digital pre-distortion method based on neural network

被引：0

作者：

Zhang, Lie ^{[1
]}

Feng, Yan ^{[1
]}

机构：

[1] Department of Electronics Engineering, Northwestern Polytechnical University, Xi'an,710129, China

来源：

Xibei Gongye Daxue Xuebao/Journal of Northwestern Polytechnical University | 2014年 / 32卷 / 06期

关键词：

Genetic algorithms - Radio frequency amplifiers - Digital radio - Neural networks - Polynomials - Convergence of numerical methods - Mean square error;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

An optimized digital predistortion (DPD) approach of radio frequency (RF) power amplifier (PA) is proposed; It employs real-valued neural network model based on low-order generalized memory polynomial, which is optimized with genetic algorithm. It cascades genetic algorithm optimized low-order generalized memory polynomial and neural network model to increase matched degree between correction model and distortion of the PA. It can not only improve the correction ability of the model but also accelerate convergence speed of the model. The 60MHz LTE 3 carrier signal is employed to do measurement. Results show that the proposed approach is 6 dB better than real-valued focused time delay neural network model in ACLR and gives faster convergence speed. ©, 2014, Northwestern Polytechnical University. All right reserved.

引用

页码：967 / 973