自适应稀疏重构的双频预失真结构  

Dual-frequency predistortion structure with adaptive sparse reconstruction

作　　者：高明明[1,2] 吴月 房少军 南敬昌[1] Gao Mingming;Wu Yue;Fang Shaojun;Nan Jingchang(School of Electronic Information&Engineering,Liaoning Technical University,Huludao Liaoning 125105,China;College of Informa-tion Science&Technology,Dalian Maritime University,Dalian Liaoning 116026,China)

机构地区：[1]辽宁工程技术大学电子与信息工程学院,辽宁葫芦岛125105 [2]大连海事大学信息科学技术学院,辽宁大连116026

出　　处：《计算机应用研究》2020年第8期2428-2432,共5页Application Research of Computers

基　　金：国家自然科学基金青年科学基金资助项目(61701211);辽宁省特聘教授项目(551710007004);辽宁省高校重点实验室项目(LJZS007)。

摘　　要：针对双频功放预失真系统采样率过高的问题,提出一种基于压缩感知的自适应稀疏预失真结构,先通过基于分段多项式模型的记忆效应补偿器,再将信号融合理解为压缩感知采样重构问题,即在预失真反馈回路,利用自适应稀疏算法高精度重构遗失的五阶及高阶交调信号,使系数权值的最小均方解逼近最优,降低采集误差提升线性化效果。实验结果表明,在提高系统稳定性的同时,NMSE显示较2D-MP、2D-CPWL提高了约2~3 dB,ACPR大约改善了20 dBc。对降低双频带预失真采样率同时提升功放线性度具有重大意义。To solve the problem of high sampling rate in the dual frequency amplifier predistortion system,this paper proposed an adaptive sparse predistortion structure based on compressed sensing.Firstly,it used the memory effect compensator based on piecewise polynomial model,and then understood the signal fusion as the compressive sensing sampling reconstruction problem.That was,the pre-distortion feedback loop used the adaptive sparse algorithm to reconstruct the missing intermodulation signals of the fifth order and the higher order with high accuracy,so that the minimum mean square solution of coefficient weight value approximated to the optimal value.This method could reduce the acquisition error and improve linearization.The experimental results show that normalized mean squared error display is about 2~3 dB better than 2 dimensional-memory polynomial(2D-MP) and 2 dimensional-canonical piecewise linear(2 D-CPWL) in improving the system stability.The adjacent channel power radio(ACPR) can be improved to 20 dBc.It is of great significance to reduce the predistortion sampling rate of dual frequency band and improve the linearity of power amplifier.

关 键 词：压缩感知 子空间追踪 预失真 双频功放 自适应重构 

分 类 号：TN722.5[电子电信—电路与系统]