帶限數(shù)字預(yù)失真系統(tǒng)關(guān)鍵技術(shù)研究
發(fā)布時間：2018-12-11 04:08
【摘要】：隨著信息容量和數(shù)據(jù)傳輸速率的日益增長,現(xiàn)代無線通信系統(tǒng)大都采用了高階的調(diào)制技術(shù)來充分利用有限的頻譜資源,同時數(shù)據(jù)傳輸速率的增長使得信號的帶寬大幅度增加。采用高階調(diào)制體制技術(shù)的信號具有功率峰均比高的特點(diǎn),這惡化了功放的高線性度和高效率的矛盾。功放的線性化技術(shù)是緩解這個矛盾的技術(shù)手段,其中,數(shù)字預(yù)失真技術(shù)以其線性化能力強(qiáng),適應(yīng)性強(qiáng)等特點(diǎn)成為了當(dāng)前的研究熱點(diǎn)。然而,隨著信號帶寬的增加,造成了反饋回路需要高采樣率ADC的問題,增加了數(shù)字預(yù)失真技術(shù)的實(shí)現(xiàn)的成本和難度。針對上述問題,本文對寬帶的數(shù)字預(yù)失真技術(shù)展開了研究。首先,由于在帶限數(shù)字預(yù)失真系統(tǒng)中,傳統(tǒng)的帶限記憶多項(xiàng)式模型需要階數(shù)非常高的帶限濾波器來保證其精度,這造成了傳統(tǒng)的帶限記憶多項(xiàng)式模型復(fù)雜度高的問題,為此提出了一種低復(fù)雜度的帶限記憶多項(xiàng)式模型,該模型可以降低所需的帶限濾波器階數(shù),而且具有較高的性能。實(shí)驗(yàn)證明,與傳統(tǒng)的帶限記憶多項(xiàng)式模型相比,在同等性能條件下,模型系數(shù)數(shù)目和浮點(diǎn)運(yùn)算數(shù)目分別下降了大約42%和65%。其次,針對帶限數(shù)字預(yù)失真系統(tǒng)中的環(huán)路延時問題,采用自適應(yīng)的算法求解小數(shù)時延,可提高算法估計(jì)的精度,并且采用自適應(yīng)算法的估算算法,對于工程實(shí)現(xiàn)具有一定的參考價值。最后,由于帶限數(shù)字預(yù)失真技術(shù)只能改善模擬濾波器帶寬內(nèi)的頻譜,超出模擬濾波器的帶寬外的頻譜并不能得到改善,這部分信息將會干擾鄰近信道,并且,帶限預(yù)失真技術(shù)需采用帶寬較窄的帶限濾波器濾除掉受到模擬濾波器滾降等邊緣不理想因素影響的頻譜的局限性問題。為此,提出一種分頻段的記憶多項(xiàng)式模型,實(shí)驗(yàn)證明該模型比傳統(tǒng)的帶限記憶多項(xiàng)式模型具有更低的復(fù)雜度,對信號的ACPR改善了19dBc,EVM改善了5.51%,同時對超出模擬濾波器的帶寬外的頻譜有一定的線性化能力。
[Abstract]:With the increasing of information capacity and data transmission rate, modern wireless communication systems mostly use high-order modulation technology to make full use of limited spectrum resources. At the same time, the increase of data transmission rate makes the signal bandwidth increase greatly. The signal with high order modulation system has the characteristics of high PAPR, which exacerbates the contradiction between high linearity and high efficiency of power amplifier. The linearization technology of power amplifier is the technical means to alleviate this contradiction. Among them, digital predistortion technology has become the focus of current research because of its strong linearization ability and strong adaptability. However, with the increase of signal bandwidth, the problem of high sampling rate ADC is caused by the feedback loop, which increases the cost and difficulty of digital predistortion technology. In order to solve the above problems, the digital predistortion technology of broadband is studied in this paper. First of all, in the band-limited digital predistortion system, the traditional band-limited memory polynomial model needs a very high order band-limited filter to ensure its accuracy, which leads to the high complexity of the traditional band-limited memory polynomial model. In this paper, a low complexity band-limited memory polynomial model is proposed, which can reduce the order of band-limited filter and has high performance. Experimental results show that compared with the traditional band-limited memory polynomial model, the number of coefficients and the number of floating-point operations are decreased by about 42% and 65% respectively under the same performance conditions. Secondly, for the loop delay problem in the band-limited digital predistortion system, adaptive algorithm is used to solve the fractional delay, which can improve the estimation accuracy of the algorithm, and the estimation algorithm of the adaptive algorithm is adopted. It has certain reference value for engineering realization. Finally, because band-limited digital predistortion technology can only improve the spectrum in the bandwidth of analog filter, the spectrum beyond the bandwidth of analog filter can not be improved. This information will interfere with adjacent channels, and, The bandlimited predistortion technique requires the use of bandlimited filters with narrow bandwidth to remove the limitation of the frequency spectrum affected by the undesired edge factors such as the roll down of the analog filter. For this reason, a memory polynomial model in frequency division band is proposed. The experimental results show that this model has lower complexity than the traditional band-limited memory polynomial model, and improves the ACPR of the signal by 5.51%. At the same time, there is a certain linearization ability for the spectrum beyond the bandwidth of the analog filter.
【學(xué)位授予單位】：中國工程物理研究院
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：TN713
本文編號：2371837
[Abstract]:With the increasing of information capacity and data transmission rate, modern wireless communication systems mostly use high-order modulation technology to make full use of limited spectrum resources. At the same time, the increase of data transmission rate makes the signal bandwidth increase greatly. The signal with high order modulation system has the characteristics of high PAPR, which exacerbates the contradiction between high linearity and high efficiency of power amplifier. The linearization technology of power amplifier is the technical means to alleviate this contradiction. Among them, digital predistortion technology has become the focus of current research because of its strong linearization ability and strong adaptability. However, with the increase of signal bandwidth, the problem of high sampling rate ADC is caused by the feedback loop, which increases the cost and difficulty of digital predistortion technology. In order to solve the above problems, the digital predistortion technology of broadband is studied in this paper. First of all, in the band-limited digital predistortion system, the traditional band-limited memory polynomial model needs a very high order band-limited filter to ensure its accuracy, which leads to the high complexity of the traditional band-limited memory polynomial model. In this paper, a low complexity band-limited memory polynomial model is proposed, which can reduce the order of band-limited filter and has high performance. Experimental results show that compared with the traditional band-limited memory polynomial model, the number of coefficients and the number of floating-point operations are decreased by about 42% and 65% respectively under the same performance conditions. Secondly, for the loop delay problem in the band-limited digital predistortion system, adaptive algorithm is used to solve the fractional delay, which can improve the estimation accuracy of the algorithm, and the estimation algorithm of the adaptive algorithm is adopted. It has certain reference value for engineering realization. Finally, because band-limited digital predistortion technology can only improve the spectrum in the bandwidth of analog filter, the spectrum beyond the bandwidth of analog filter can not be improved. This information will interfere with adjacent channels, and, The bandlimited predistortion technique requires the use of bandlimited filters with narrow bandwidth to remove the limitation of the frequency spectrum affected by the undesired edge factors such as the roll down of the analog filter. For this reason, a memory polynomial model in frequency division band is proposed. The experimental results show that this model has lower complexity than the traditional band-limited memory polynomial model, and improves the ACPR of the signal by 5.51%. At the same time, there is a certain linearization ability for the spectrum beyond the bandwidth of the analog filter.
【學(xué)位授予單位】：中國工程物理研究院
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：TN713
【參考文獻(xiàn)】
相關(guān)博士學(xué)位論文 前1條
1 李明玉;寬帶通信中功率放大器行為模型與預(yù)失真技術(shù)研究[D];電子科技大學(xué);2010年
，本文編號：2371837
本文鏈接：http://sikaile.net/kejilunwen/dianzigongchenglunwen/2371837.html
最近更新
教材專著
