【摘要】：隨著通信頻帶的擴(kuò)展,通信系統(tǒng)對功率放大器在線性度和效率上提出了更高的要求。數(shù)字預(yù)失真技術(shù)是補(bǔ)償寬帶功率放大器非線性特性和記憶效應(yīng)最有效的方式。近年來,寬帶功率放大器的數(shù)字預(yù)失真技術(shù)受到了國內(nèi)外學(xué)者的廣泛關(guān)注,開展針對其特性的預(yù)失真技術(shù)研究迫在眉睫。本文主要研究適用于寬帶功率放大器的功放建模和預(yù)失真技術(shù)。首先介紹了功率放大器行為模型,重點(diǎn)介紹了正交多項(xiàng)式模型及其對強(qiáng)記憶深度寬帶功放模型的改善作用;其次介紹了功放數(shù)字預(yù)失真技術(shù),著重介紹了預(yù)失真器的參數(shù)提取算法,包括最小二乘法(Least Squares,LS)、遞歸最小二乘法(Recursive Least Squares,RLS)、最小均方算法(Least Mean Squares,LMS)及其改進(jìn)算法,首次提出并分析了輸入矩陣條件數(shù)對RLS算法的影響,證明了使用基于正交多項(xiàng)式模型的功放預(yù)失真器,進(jìn)行RLS參數(shù)估計(jì),可避免因?qū)拵ЧΨ艔?qiáng)記憶效應(yīng)導(dǎo)致的輸入矩陣相關(guān)性上升問題,提高了建模精度;然后討論了 RLS-LMS結(jié)合算法,提出了一種基于多模型數(shù)據(jù)融合的新型參數(shù)提取算法,將收斂后的RLS解作為靜態(tài)參數(shù),LMS解作為動態(tài)參數(shù),利用最大似然函數(shù)實(shí)現(xiàn)數(shù)據(jù)融合,最終輸出適用于寬帶功放預(yù)失真器的高精度參數(shù)估計(jì)值。仿真結(jié)果表明,均方根誤差(Root Mean Square Error,RMSE)降低了 4.2dB,融合算法有效提高了預(yù)失真器的收斂速度和精度;文章最后介紹了基于正交多項(xiàng)式模型和數(shù)據(jù)融合參數(shù)提取算法的寬帶功放數(shù)字預(yù)失真系統(tǒng)方案,對其進(jìn)行仿真測試,驗(yàn)證了其有效性,并最終通過實(shí)驗(yàn)驗(yàn)證平臺對其進(jìn)行了性能測試,實(shí)測激勵信號為100MHz帶寬OFDM信號,功放為工作頻率1.95GHz的Doherty寬帶功放,實(shí)測結(jié)果表明使用常規(guī)預(yù)失真方案,功放輸出ACPR可從-31.29dBc改善至-41.16dBc,而使用新型預(yù)失真方案,可將功放輸出ACPR進(jìn)一步改善至-44.37dBc,驗(yàn)證了該方案的有效性。
[Abstract]:With the expansion of the communication frequency band, the communication system requires higher linearity and efficiency of the power amplifier. Digital predistortion is the most effective way to compensate for the nonlinear characteristics and memory effects of broadband power amplifiers. In recent years, the digital predistortion technology of wideband power amplifier has received extensive attention from scholars at home and abroad. This paper mainly studies the power amplifier modeling and predistortion technology suitable for wideband power amplifier. First of all, the behavior model of power amplifier is introduced, the orthogonal polynomial model and its improvement to the broadband power amplifier model of strong memory depth are introduced, and then the digital predistortion technology of power amplifier is introduced. The parameter extraction algorithms of predistortors, including least Squares LS, Recursive least Squares RLS, least mean Square LMS and their improved algorithms, are introduced. The influence of the number of input matrix conditions on the RLS algorithm is presented and analyzed for the first time. It is proved that using the power amplifier predistorter based on orthogonal polynomial model to estimate the RLS parameters can avoid the problem of increasing the correlation of input matrix caused by the strong memory effect of wideband power amplifier and improve the modeling accuracy. Then the RLS-LMS algorithm is discussed, and a new parameter extraction algorithm based on multi-model data fusion is proposed. The convergent RLS solution is taken as the static parameter and the LMS solution is taken as the dynamic parameter, and the maximum likelihood function is used to realize the data fusion. The final output is suitable for the high precision parameter estimation of the wideband power amplifier predistorter. The simulation results show that the root mean square error (RMSE) reduces 4.2 dB, and the convergence speed and accuracy of the predistorter are improved effectively by the fusion algorithm. Finally, the paper introduces the scheme of wideband power amplifier digital predistortion system based on orthogonal polynomial model and data fusion parameter extraction algorithm. Finally, the performance is tested by the experimental verification platform. The measured excitation signal is 100MHz bandwidth OFDM signal and the power amplifier is Doherty wideband power amplifier with 1.95GHz operating frequency. The experimental results show that the conventional predistortion scheme is used. The power amplifier output ACPR can be improved from -31.29dBc to -41.16dBc.Using a new predistortion scheme, the output ACPR can be further improved to -44.37dBc. the validity of the scheme is verified.
【學(xué)位授予單位】：南京理工大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TN722.75

【參考文獻(xiàn)】

相關(guān)期刊論文 前4條

1 曾曉洋;黎明;李志宏;陳兢;楊玉超;黃如;;微納集成電路和新型混合集成技術(shù)[J];中國科學(xué):信息科學(xué);2016年08期

2 南敬昌;李厚儒;方楊;;面向TD-LTE通信系統(tǒng)的功放新模型[J];計(jì)算機(jī)工程與應(yīng)用;2015年05期

3 吳燕鳴;孟維曉;;數(shù)字基帶預(yù)失真自適應(yīng)算法改進(jìn)與性能分析[J];移動通信;2008年24期

4 席裕庚，柴天佑，惲為民;遺傳算法綜述[J];控制理論與應(yīng)用;1996年06期

，

本文編號：2136952