本文選題：正交頻分復(fù)用 + 多用戶系統(tǒng)�。� 參考：《復(fù)旦大學(xué)》2014年博士論文

【摘要】：正交頻分復(fù)用(OFDM)技術(shù)具有良好的抗多徑干擾能力以及很高的頻譜效率,現(xiàn)已在無線通信系統(tǒng)中得到廣泛應(yīng)用。但與單載波系統(tǒng)相比,OFDM系統(tǒng)對頻偏誤差比較敏感：載波頻偏與采樣頻偏都會破壞系統(tǒng)的正交性,從而引起子載波間干擾(ICI)。而在多用戶OFDM系統(tǒng)中,不同用戶的信號經(jīng)過無線信道傳播后在接收端相互疊加,又引入了其它用戶的ICI,使得系統(tǒng)性能進(jìn)一步下降。本文考慮了OFDM系統(tǒng)中載波頻偏和采樣頻偏的估計(jì)問題,先后討論了單用戶系統(tǒng)中初始頻偏的捕捉以及多用戶系統(tǒng)中殘留頻偏的跟蹤,最后利用載波頻偏估計(jì)的思想解決了相位模糊的問題。本文首先討論了整數(shù)載波頻偏的估計(jì)�，F(xiàn)有文獻(xiàn)在估計(jì)整數(shù)頻偏時(shí)假設(shè)小數(shù)頻偏已經(jīng)被事先完全補(bǔ)償。但是實(shí)際中,由于噪聲的緣故,小數(shù)頻偏的估計(jì)會有一定的誤差,所以在補(bǔ)償之后系統(tǒng)中仍然會有一些殘留小數(shù)頻偏,并會對整數(shù)頻偏的估計(jì)產(chǎn)生較大的影響。本文分析了傳統(tǒng)算法有殘留小數(shù)頻偏時(shí)性能下降的原因,并提出了一種魯棒的整數(shù)頻偏與信道的聯(lián)合ML估計(jì)算法。仿真結(jié)果表明,本文算法的性能優(yōu)于傳統(tǒng)算法。對于小數(shù)載波頻偏與采樣頻偏,現(xiàn)有文獻(xiàn)中的聯(lián)合ML估計(jì)算法具有較高的估計(jì)精度,但是需要進(jìn)行二維窮舉式搜索,計(jì)算量較大。本文提出了一種低復(fù)雜度的小數(shù)頻偏與采樣頻偏的聯(lián)合ML估計(jì)算法：該算法首先證明小數(shù)頻偏可以直接求解,從而將原來的二維搜索變?yōu)橐痪S搜索；進(jìn)而在二階泰勒展開后,提出了一種對采樣頻偏的近似ML估計(jì)。仿真結(jié)果表明,本文算法的性能與傳統(tǒng)算法幾乎一樣,但計(jì)算量大大降低。針對多用戶OFDM系統(tǒng),由于各個(gè)用戶分別使用不同的晶振,所以頻偏誤差各不相同,接收端需要同時(shí)估計(jì)多個(gè)載波頻偏和采樣頻偏�，F(xiàn)有文獻(xiàn)對多用戶頻偏的估計(jì)采用了迫零準(zhǔn)則,因而有噪聲增強(qiáng)問題。本文提出了一種新的多用戶OFDM系統(tǒng)中載波頻偏與采樣頻偏聯(lián)合估計(jì)算法。新算法利用QR分解完成用戶分離,可以消除已估用戶對剩余用戶的干擾。為了進(jìn)一步提高估計(jì)精度,新算法還對接收信號進(jìn)行了加權(quán)平均。仿真結(jié)果表明,新算法的性能要優(yōu)于現(xiàn)有文獻(xiàn)中的算法�？紤]到晶振的溫漂會使得頻偏隨時(shí)間發(fā)生變化,本文還進(jìn)一步討論了時(shí)變多用戶系統(tǒng)中載波頻偏與信道的聯(lián)合跟蹤問題：將頻偏作為隱變量,采用EM算法求解最大后驗(yàn)估計(jì)；為了能夠求得E步中隱變量的條件期望,本文利用變分推斷來近似原來的后驗(yàn)概率。仿真結(jié)果表明,本文算法可以在較少的迭代次數(shù)下達(dá)到較高的跟蹤精度。最后,本文利用載波頻偏估計(jì)的思想解決了相位模糊的問題。在跟蹤頻偏與信道變化的過程中,需要大量的導(dǎo)頻開銷。盲信道估計(jì)不依賴導(dǎo)頻信號,但辨識結(jié)果存在相位模糊的問題。最近的研究表明,相位模糊是由于以前的算法忽略了發(fā)送信號中先驗(yàn)的星座圖信息。本文采用載波頻偏估計(jì)的思想,證明了模糊相位可以分解為小數(shù)相位和整數(shù)相位兩部分,而且它們的可辨識性也與小數(shù)頻偏、整數(shù)頻偏相類似。該結(jié)論可以進(jìn)一步拓展到有任意多個(gè)星座圖同時(shí)使用的情況。仿真結(jié)果表明,本文算法的性能與利用導(dǎo)頻消除相位模糊的性能不差上下。
[Abstract]:Orthogonal frequency division multiplexing (OFDM) technology has good anti multipath interference ability and high spectral efficiency. It has been widely used in wireless communication system. But compared with single carrier system, OFDM system is more sensitive to frequency offset error: carrier frequency offset and sampling frequency offset can break the orthogonality of the system and cause inter carrier interference. (ICI). In the multiuser OFDM system, the signal of different users superimposition each other at the receiving end after the wireless channel is propagated and the ICI of other users is introduced. The performance of the system is further reduced. This paper considers the estimation of the carrier frequency offset and sampling frequency offset in the OFDM system, and discusses the capture of the initial frequency offset in the single user system. As well as the tracking of residual frequency offset in multiuser system, the problem of phase ambiguity is solved by using the idea of carrier frequency offset estimation. First, the estimation of integer carrier frequency offset is discussed. There will be some error in the estimation, so there will still be some residual decimal frequency offset in the system after compensation, and it will have a great influence on the estimation of integer frequency offset. In this paper, the reason for the performance decline of the traditional algorithm with residual decimal frequency offset is analyzed, and a robust combined ML estimation algorithm for the integer frequency offset and the channel is proposed. The real results show that the performance of the proposed algorithm is superior to the traditional algorithm. For the decimal carrier frequency offset and sampling frequency offset, the combined ML estimation algorithm in the existing literature has a high estimation accuracy, but it needs a two-dimensional exhaustive search and a large amount of computation. This paper presents a combined ML estimation of the low complex number frequency offset and sampling frequency offset. Algorithm: this algorithm first proves that the frequency offset of the decimal can be solved directly, and then turns the original two-dimensional search into one dimension search. Then, after two order Taylor expansion, an approximate ML estimation for sampling frequency offset is proposed. The simulation results show that the performance of this algorithm is similar to that of the traditional algorithm, but the computation is greatly reduced. OFDM system, because each user uses the different crystal oscillator respectively, the frequency offset error is different, the receiver needs to estimate multiple carrier frequency offset and sampling frequency offset at the same time. The existing literature has adopted the zero forcing criterion for the estimation of multiuser frequency offset. Therefore, there is a noise enhancement problem. In this paper, a new carrier frequency offset in the multi-user OFDM system is proposed. The new algorithm uses the sampling frequency offset estimation algorithm. The new algorithm uses the QR decomposition to complete the user separation, which can eliminate the interference of the estimated user to the remaining users. In order to further improve the estimation precision, the new algorithm also carries out weighted average of the received signal. The simulation results show that the performance of the new algorithm is better than the existing algorithms. Temperature drift will change frequency offset with time. This paper further discusses the joint tracking problem of carrier frequency offset and channel in time-varying multiuser systems: using frequency offset as implicit variable and using EM algorithm to solve the maximum posterior estimation. In order to find the conditional expectation of hidden variables in E step, this paper uses variational inference to approximate the original The simulation results show that this algorithm can achieve higher tracking accuracy under less iterative times. Finally, the problem of phase ambiguity is solved by the idea of carrier frequency offset estimation. In the process of tracking frequency offset and channel change, a large number of pilot open pins are needed. Blind channel estimation is not dependent on pilot signal, but identification is not dependent on the pilot signal. The result shows that there is a problem of phase ambiguity. The recent study shows that the phase ambiguity is due to the previous algorithm neglecting the prior constellation information in the transmitted signal. In this paper, the idea of carrier frequency offset estimation proves that the fuzzy phase can be decomposed into two parts of the decimal phase and the integer phase, and their identifiability is also associated with the decimal frequency offset. The integer frequency offset is similar. This conclusion can be further extended to the situation that any multiple constellation diagrams are used simultaneously. The simulation results show that the performance of the proposed algorithm is not very different from the performance of using pilot to eliminate phase ambiguity.
【學(xué)位授予單位】：復(fù)旦大學(xué)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2014
【分類號】：TN929.53

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 薛斌;安建平;汪靜;田甜;;一種直接序列擴(kuò)頻系統(tǒng)的大頻偏二次捕獲算法[J];北京理工大學(xué)學(xué)報(bào);2011年11期

2 李新民;王旭亮;扈平;;基于頻偏預(yù)校正匹配濾波的同步方法[J];無線電工程;2012年02期

3 孫金榮;;調(diào)頻電視的過頻偏傳輸[J];電視技術(shù);1987年06期

4 Ю.Д.Болмусов;杰雄;;信號發(fā)生器微小寄生頻偏的測量[J];國外計(jì)量;1988年06期

5 王景明;;兩種精確測量調(diào)頻頻偏的新方法[J];國外電子測量技術(shù);2008年03期

6 馮祥;梁偉洋;;頻偏穩(wěn)健的分級調(diào)制分類算法[J];電訊技術(shù);2011年07期

7 王瑞斌;;調(diào)幅度頻偏容限測量原理[J];內(nèi)蒙古廣播與電視技術(shù);2005年04期

8 秦勇;張邦寧;郭道省;;大頻偏下直接序列擴(kuò)頻系統(tǒng)的碼捕獲技術(shù)[J];無線通信技術(shù);2005年04期

9 張公禮;張凱飛;羅宏杰;;一種糾正殘余頻偏的均方誤差反饋算法[J];計(jì)算機(jī)仿真;2008年05期

10 王雪;戴伏生;李尚;;基于數(shù)字化插值的頻偏抑制因子的計(jì)算[J];信息技術(shù);2013年12期

相關(guān)會議論文 前6條

1 李穎;張琛;魏急波;;相干QPSK/QAM OFDM系統(tǒng)中的殘余頻偏補(bǔ)償算法[A];通信理論與信號處理新進(jìn)展——2005年通信理論與信號處理年會論文集[C];2005年

2 蔡勇華;蘇e，

本文編號：2055085