首页期刊简介编委会征稿启事出版道德声明审稿流程读者订阅论文查重联系我们English
引用本文
  • 袁江南,唐 骏,伊锦旺.一种级联两箱通用数字预失真器及其辨识算法[J].电讯技术,2018,58(1): - .    [点击复制]
  • YUAN Jiangnan,TANG Jun,YIN Jinwang.A cascaded versatile digital predistorter and its identification algorithm[J].,2018,58(1): - .   [点击复制]
【打印本页】 【下载PDF全文】 查看/发表评论下载PDF阅读器关闭

←前一篇|后一篇→

过刊浏览    高级检索

本文已被:浏览 1871次   下载 430 本文二维码信息
码上扫一扫!
一种级联两箱通用数字预失真器及其辨识算法
袁江南,唐骏,伊锦旺
0
(厦门理工学院 光电与通信工程学院,福建 厦门 361024)
摘要:
提出了一种由单形规范线性分段(SCPWL)函数与记忆多项式级联的数字预失真器,并给出了复数域两步最小二乘参数辨识算法。不同于以往一种预失真器适用一种功放模型的情况,所提的预失真算法利用SCPWL函数的分段特性以及记忆多项式的非线性记忆特性,在完成参数辨识的同时自动地调整结构,可适用于传统以及强非线性新型功放模型的线性化补偿。将所提预失真器分别应用于传统记忆多项式、两箱模型以及新型包络跟踪功放。经过计算机仿真,功放输出的幅频特性和频谱曲线表明所提预失真器能够有效地补偿多种功放的非线性特性。算法仿真比较结果也表明,针对包络跟踪功放,所提复数两步最小二乘算法的邻道泄漏比(ACLR)可改善约35 dB,性能优于最小均方(LMS)类算法约30 dB。
关键词:  数字预失真  级联两箱模型  SCPWL函数  记忆多项式  包络跟踪
DOI:
基金项目:国家自然科学基金资助项目(61701422);福建省自然科学基金资助项目(2015J01670);厦门市科技计划项目(3502Z20153017)
A cascaded versatile digital predistorter and its identification algorithm
YUAN Jiangnan,TANG Jun,YIN Jinwang
()
Abstract:
A digital predistorter of cascaded simplicial canonical piecewise linear function(SCPWL) and memory polynomial is proposed,and the complex field two-step least square(TSLS) parameter identification algorithm is presented as well. Different from the case that one type of predistorter adapts to specially appointed amplifier,the proposed predistortion algorithm is able to adjust its structure during the parameter identification period automatically to accommodate multiple amplifier models by using the piecewise features of SCPWL function and the nonlinear memory characteristics of memory polynomial. The proposed predistorter algorithm has been used to compensate for traditional polynomial,two-box model and state-of-art envelope tracking amplifiers. Amplitude-frequency and spectrum curves of computer simulations show that the proposed predistorter can adaptively compensate for all these amplifiers’ nonlinear characteristics. Algorithm comparison also shows that,in case of envelope tracking amplifiers,the adjacent channel leakage ratio(ACLR) performance of TSLS algorithm can be improved over 35 dB approximately,superior to that of LMS class algorithms about 30 dB.
Key words:  digital predistortion  cascaded two box model  scpwl function  memory polynomial  envelope tracking
安全联盟站长平台