一种基于记忆LUT的多倍频程预失真方案

发布时间：2018-03-11 18:14

本文选题：行为模型　切入点：预失真　出处：《科学技术与工程》2014年34期 　论文类型：期刊论文

【摘要】：针对多载波无线通信系统中多倍频程功率放大器存在的强非线性失真与记忆效应失真问题,提出了一种基于记忆查找表的LUT+MP多倍频程预失真方案。方案模型由记忆查找表(LUT)和记忆多项式(MP)串联构成,采用间接学习法的LMS和RLS算法对模型参数训练辨识。仿真结果表明,针对维纳模型高功率放大器,在输出功率回退OBO=2.27 d B时,对Wiener模型功放的三阶互调失真改善达到了55.9 d B,五阶互调失真改善达到了44.5 d B。与现有的LUT+FIR和LUT+并联FIR预失真方案相比较,能带来更好的建模精度和线性化效果。同时,设计使用的学习算法简单实用,稳定性能好。
[Abstract]:In order to solve the problem of strong nonlinear distortion and memory distortion in multi-frequency doubly-path power amplifier in multicarrier wireless communication system, In this paper, a multi-octave path predistortion scheme of LUT MP based on memory lookup table is proposed. The scheme model consists of memory lookup table (LUTT) and memory polynomial (MPP) series. The LMS and RLS algorithms of indirect learning method are used to train and identify the model parameters. The simulation results show that for Wiener model high power amplifier, when the output power is back to OBO=2.27 dB, The improvement of third-order Intermodulation distortion and five-order Intermodulation distortion for Wiener power amplifier is 55.9 dB and 44.5 dB respectively. Compared with the existing LUT FIR and LUT parallel FIR predistortion schemes, it can bring better modeling accuracy and linearization effect. The learning algorithm designed and used is simple and practical with good stability.
【作者单位】：南昌大学信息工程学院电子系;
【分类号】：TN913.6;TN722.75

【相似文献】