1、MIMOOFDM系統(tǒng)中信道估計(jì)及最優(yōu)導(dǎo)頻序列設(shè)計(jì)的研究.txt只要你要，只要我有，你還外邊轉(zhuǎn)什么阿老實(shí)在我身邊待著就行了。聽我的就是,問那么多干嘛，我在你身邊,你還走錯(cuò)路!跟著我!不能給你幸福是我的錯(cuò)，但誰讓你不幸福，我TMD去砍了他分類號:UDC:TN 929.531621.3學(xué)號039072東博d卜南大學(xué)學(xué)位論文MIMO OFDM系統(tǒng)中信道估計(jì)及最優(yōu) 導(dǎo)頻序列設(shè)計(jì)的研究研究生姓名導(dǎo)師姓名:胡蝶楊綠溪教授申請學(xué)位級別工學(xué)博士學(xué)科專業(yè)名稱論文提交日期2006年6月26日論文答辯日期學(xué)位授予單位東南大學(xué)學(xué)位授予日期答辯委員會主席鄂廣增評閱人2006年月日2006年6月5日 CHANNEL EST

2、IMATION ANDOPTIMAL PILOT SEQUENCE DESIGN FOR MIMO OFDM SYSTEMSA Dissertation Submitted toSoutheast UniversityFor the Academic Degree of Doctor of EngineeringHU DieSupervised byProf. YANG Lu-xiSchool of Information Science and Engineering Southeast UniversityJune 2006東南大學(xué)學(xué)位論文獨(dú)創(chuàng)性聲明 本人聲明所呈交的學(xué)位論文是我個(gè)人在導(dǎo)師

3、指導(dǎo)下進(jìn)行的研究工作及取得的研究成果。盡我所知，除了文中特別加以標(biāo)注和致謝的地方外，論文中不包含其他人已經(jīng)發(fā)表或撰寫過的研究成果，也不包含為獲得東南大學(xué)或其它教育機(jī)構(gòu)的學(xué)位或證書而使用過的材料。與我一同工作的同志對本研究所做的任何貢獻(xiàn)均已在論文中作了明確的說明并表示了謝意。研究生簽名:日期:東南大學(xué)學(xué)位論文使用授權(quán)聲明 東南大學(xué)、中國科學(xué)技術(shù)信息研究所、國家圖書館有權(quán)保留本人所送交學(xué)位論文的復(fù)印件和電子文檔，可以采用影印、縮印或其他復(fù)制手段保存論文。本人電子文檔的內(nèi)容和紙質(zhì)論文的內(nèi)容相一致。除在保密期內(nèi)的保密論文外，允許論文被查閱和借閱，可以公布(包括刊登)論文的全部或部分內(nèi)容。論文的公布(包

4、括刊登)授權(quán)東南大學(xué)研究生院辦理。研究生簽名:導(dǎo)師簽名:日期:摘要MIMO OFDM系統(tǒng)中信道估計(jì)及最優(yōu)導(dǎo)頻序列設(shè)計(jì)的研究摘要 正交頻分復(fù)用(Orthogonal Frequency Division Multiplexing一OFDM)是信息高速傳輸?shù)挠行Ъ夹g(shù)，由于它可以抵抗信道的多徑衰落，且具有頻譜利用率高、實(shí)現(xiàn)簡單等一系列優(yōu)點(diǎn)，因此成為了未來移動通信系統(tǒng)的主要候選方案之一。信道估計(jì)是OFDM系統(tǒng)中的關(guān)鍵技術(shù)之一，利用信道信息的相干解調(diào)系統(tǒng)與不利用信道信息的差分解調(diào)系統(tǒng)相比，在信噪比(signal-to-noise ratio, SNR)上可以有3 dB以上的性能提高。因此在要求高傳輸速率

5、、高傳輸質(zhì)量和高移動性的4G (FourthGeneration)通信系統(tǒng)中，信道估計(jì)是不可缺少的一環(huán)。本論文主要針對這個(gè)關(guān)鍵技術(shù)及相關(guān)的最優(yōu)導(dǎo)頻序列設(shè)計(jì)問題進(jìn)行了研究。 由于信道估計(jì)已得到廣泛的研究，因此國際和國內(nèi)已有較多的相關(guān)技術(shù)文獻(xiàn)。本論文對其中一些經(jīng)典和常用的信道估計(jì)算法進(jìn)行了總結(jié)，然后在此基礎(chǔ)上對多輸入多輸出(Multiple Input MultipleOutput, MIMO) OFDM系統(tǒng)進(jìn)行了分析，得到了一些新的結(jié)論。針對MIMO OFDM系統(tǒng)中進(jìn)行信道時(shí)域響應(yīng)估計(jì)時(shí)有可能出現(xiàn)的方程欠定問題，論文提出了一種聯(lián)合相繼發(fā)送的導(dǎo)頻序列進(jìn)行共同估計(jì)的改進(jìn)算法。當(dāng)發(fā)送的導(dǎo)頻序列較為充足

6、時(shí)，可以得到信道響應(yīng)的若干個(gè)估計(jì)，根據(jù)每個(gè)估計(jì)的均方誤差(Mean Square Error, MSE)，論文給出了一種結(jié)合方式以獲得最終較優(yōu)的信道響應(yīng)估計(jì)。 實(shí)際的OFDM系統(tǒng)通常會使用虛擬子載波，以此來避免發(fā)送信號受到發(fā)送端成形低通濾波器的影響。在設(shè)計(jì)導(dǎo)頻序列時(shí)，若考慮系統(tǒng)的虛擬子載波，則在某些情況下，傳統(tǒng)均勻分布的最優(yōu)導(dǎo)頻序列不再可用，因此如何設(shè)計(jì)非均勻分布的最優(yōu)導(dǎo)頻序列就顯得十分必要。論文對這個(gè)問題進(jìn)行了深入研究，在基于信道時(shí)域響應(yīng)最小二乘(Least Square, LS)估計(jì)算法的基礎(chǔ)上，依據(jù)使估計(jì)均方誤差最小的準(zhǔn)則推導(dǎo)了MIMO OFDM系統(tǒng)中非均勻分布下的最優(yōu)導(dǎo)頻序列。經(jīng)過推導(dǎo)

7、可以證明，此時(shí)導(dǎo)頻序列的設(shè)計(jì)與求解一個(gè)約束不等式方程組是等效的。對于這個(gè)約束不等式方程組，論文設(shè)計(jì)了如下三種求解算法來獲得所需的最優(yōu)導(dǎo)頻，分別是:求解一個(gè)等效的約束規(guī)劃問題;求解一個(gè)等效的非約束優(yōu)化問題;和利用Gram-Schmidt正交化進(jìn)行直接求解。 通常情況下，信道可以等效為一個(gè)有限抽頭的FIR濾波器，但對于稀疏信道，利用參數(shù)化模型則可以更準(zhǔn)確地逼近真實(shí)情況。由于參數(shù)化模型只包含信道路徑時(shí)延和路徑增益兩種參數(shù)，因此在信道稀疏時(shí)可以降低估計(jì)參數(shù)的維數(shù)，從而提高估計(jì)的性能。論文基于這種參數(shù)化信道模型，提出了一種針對稀疏信道的最大似然(Maximum Likelihood, ML)估計(jì)算法。該

8、算法首先沿用已有的最小描述長度(Minimum Description Length, MDL)準(zhǔn)則來檢測信道中的路徑數(shù)，然后利用估計(jì)信道參數(shù)的旋轉(zhuǎn)不變法(Estimation of Signal Parameters by Rotational Invariance Techniques, ESPRIT)來獲得路徑時(shí)延的初始估計(jì)。最后，利用最大似然方法對時(shí)延進(jìn)行跟蹤，對路徑增益進(jìn)行估計(jì)。為了降低多參數(shù)ML估計(jì)問題的復(fù)雜度，論文利用廣義交互空間期望最大化(Space-Alternating GeneralizedExpectation-Maximization, SAGE)算法將其分解為一系列

9、兩參數(shù)的ML估計(jì)問題，通過迭代得到最終解。對于每一次迭代中，某一信道路徑上時(shí)延與增益的兩參數(shù)ML估計(jì)問題，論文分別采用一種基于牛頓迭代的快速M(fèi)L算法和一種簡單的求和平均方法來獲得。 子載波間干擾(Intercarrier Interference, ICI)會破壞子載波之間的正交性，從而惡化OFDM系統(tǒng)的性能。由于消除ICI需要利用到信道信息，因此信道估計(jì)對于快時(shí)變環(huán)境下的OFDM系統(tǒng)來說東南大學(xué)博士學(xué)位論文是十分關(guān)鍵的。論文對這個(gè)問題進(jìn)行了深入研究，針對不同的時(shí)變環(huán)境，提出了如下三種時(shí)變的OFDM信道估計(jì)算法，分別為:(1)基于低通插值的估計(jì)算法。該算法首先利用導(dǎo)頻序列估計(jì)出一個(gè)OFDM符號

10、間隔內(nèi)給定時(shí)刻上的信道時(shí)域沖擊響應(yīng)，然后通過對前后若干個(gè)給定時(shí)刻上已估計(jì)出的信道響應(yīng)進(jìn)行低通插值，來獲得所有時(shí)刻上的信道時(shí)域響應(yīng)。(2)基于分段線性模型的估計(jì)算法。該算法對時(shí)變信道進(jìn)行線性建模，然后通過估計(jì)線性模型的斜率和參考點(diǎn)值來獲得所有的待估信道響應(yīng)。該算法適用于系統(tǒng)相對多譜勒頻率小于0.1的情況。(3)基于基擴(kuò)展模型的估計(jì)算法。與前一種算法的思想類似，該算法也是對時(shí)變信道進(jìn)行建模，然后通過估計(jì)模型參數(shù)來獲得所有待估的時(shí)變信道響應(yīng)。由于基擴(kuò)展模型與Jakes模型的擬合程度較高，因此該算法可以適用于系統(tǒng)相對多譜勒頻率較大的情況。導(dǎo)頻序列的設(shè)計(jì)對信道估計(jì)算法的性能有著重要影響，好的導(dǎo)頻序列能極

11、大改善估計(jì)算法的性能。因此，論文在給出上述凡種時(shí)變信道估計(jì)算法的同時(shí)，還分別推導(dǎo)了相應(yīng)于這些算法最優(yōu)或次優(yōu)的導(dǎo)頻序列。關(guān)鍵詞:MIMO通信、正交頻分復(fù)用、多徑信道、時(shí)變信道、信道估計(jì)、導(dǎo)頻序列AbstractChannel Estimation and Optimal Pilot Sequence Design forMIMO OFDM SystemsAbstract Orthogonal Frequency Division Multiplexing (OFDM) is an effective method for high data ratetransmission, and is be

12、ing pursued as a potential candidate for the future mobile communication systemsdue to its robustness against frequency-selective fading channels, bandwidth efficiency and simpleimplementation. Channel estimation is one of the key techniques for OFDM systems. Compared with thedifferential demodulati

13、onn, coherent demodulation with accurate channel information can achieve 3-4 dBgain in signal-to-noise ratio (SN助.Therefore, in the fourth generation (4G) systems which require highdata rate, high quality and high mobility, channel estimation is an indispensable part. This thesis focuses onthis topi

14、c and the corresponding design issue of the optimal pilot sequence. Since channel estimation has attracted a lot of attention, there have been many literatures on this issue.This thesis summarizes some classical channel estimation methods. After analyzing the Multiple InputMultiple Output (MIMO) OFD

15、M systems, some new conclusions are then obtained. In order to avoidsolving underdetermined equations for MIMO OFDM systems, an improvement algorithm which combinessuccessively transmitted pilot sequences is proposed. If pilot sequences are sufficient, more than oneestimate can be obtained. Conseque

16、ntly, a combined way with consideration of the estimation mean squareerror (MSE) is given. In practical OFDM systems, virtual subcarriers are usually used to avoid the transmitted data beingdistorted by the low-pass filter on the transmitter side. Thus, with the consideration of these virtualsubcarr

17、iers, we can find that the conventional uniformly spaced approach will not be applicable anymore insome situations. It is, therefore, necessary to derive the optimal pilot sequences with the nonuniform pilottone placement. Based on the least square (LS) channel estimation method, the optimal pilot s

18、equences ofMIMO OFDM systems are derived in this thesis to minimize the estimation MSE. It can be shown thatsuch a design issue is equivalent to solving a system of linear inequalities. For such inequalities, threesolving algorithms are given, which are solving an equivalent constrained programming

19、problem, solvingan equivalent unconstrained optimization problem, and directly using Gram-Schmidt orthogonalization,respectively. Usual玫channel can be modeled as a finite impulse response (FI助filter. But for sparse multipathchannel, the parametric model can approximate the actual environment better.

20、 Since there are only pathdelays and path gains to be estimated in the parametric model, the estimation dimension can be reduced alot for sparse channel, and consequently, the estimation performance can be improve. Based on parametricchannel model, a maximum likelihood (ML) estimation scheme is prop

21、osed for sparse channel in this thesis.First, the number of channel path are detected卜minimum-description-length (MDL) criterion and channelpath delays are initialized by Estimation of signal-parameters-卜-rotational-invariance-techniques(ESPRIT). Then the path delays and path gains are obtained by p

22、roposed ML approach. To reduce theTTT東南大學(xué)博士學(xué)位論文complexity of multi-parameter ML estimation, the space-alternating generalized expectation-maximization(SAGE) algorithm is applied to decouple the multi-parameter ML problem into some separatetwo-parameter ML problems. And the final channel estimates ar

23、e obtained卜iteration. For eachtwo-parameter ML problem, a fast ML algorithm based on Newton's method and a simple averaged methodare employed to estimate the path delay and path gain, respectively. Intercarrier interference (ICI) will destroy the orthogonality of subcarriers, which worsens the O

24、FDMsystem performance. Sine mitigating ICI needs channel information, estimating rapidly time-varyingchannel is very important. For different time-varying environments, three channel estimation schemes areproposed, which are as follows. (1) The method based on low-pass interpolation. After the chann

25、el impulseresponses (CI助at the given times are estimated by pilot sequence, the CIR at the remained times can beobtained by low-pass interpolation. (2) The method based on piece-wise linear model. Since the channelduring one OFDM symbol period is modeled as a linear function, the CIR can be obtained

26、 by onlyestimating the slope and the reference signal of the linear model. This method is suitable for systems wherethe relative Doppler frequency smaller than 0.1. (3) The method based on basis expansion model (BEM).Similar to the former method, this method also models the time-varying channel, and

27、 then obtains the CIRthrough estimating the model parameters. Since BEM can well approximate the Jakes' model, this methodis suitable for systems where the relative Doppler frequency is rather high. Pilot sequence design plays animportant role in channel estimation. Therefore, in this thesis, th

28、e optimal pilot sequences are also derivedalong with the proposed channel estimation schemes mentioned above.Key Words: MIMO communications, OFDM, multipath fading channel, time-varying channel, channelestimation, pilot sequenceTV目錄目錄摘要.IAbstract.III符號定義. VII縮略詞.VIII第一章緒論.1 1.1引言.1 1. 2 OFDM技術(shù)的歷史與發(fā)展

29、. 2 1. 3 OFDM的基本原理及特點(diǎn). 2 1.4論文的主要工作和內(nèi)容安排. 4第二章OFDM系統(tǒng)中基于導(dǎo)頻序列的信道估計(jì). 8 2.1引言. 8 2. 2 OFDM系統(tǒng)等效基帶模型. 9 2. 3 OFDM系統(tǒng)中基于導(dǎo)頻序列的經(jīng)典信道估計(jì).10 2.3.1信道頻域響應(yīng)估計(jì)算法.11 2. 3.2信道時(shí)域響應(yīng)估計(jì)算法.13 2.4 MIMO OFDM系統(tǒng)中基于導(dǎo)頻序列的信道估計(jì).14 2.5一種改進(jìn)的MIMO OFDM系統(tǒng)信道時(shí)域響應(yīng)估計(jì)算法.17 2.6仿真結(jié)果及討論. 20 2.7本章小結(jié). 21第三章MIMO OFDM系統(tǒng)中基于非均勻?qū)ьl分布的最優(yōu)導(dǎo)頻序列的設(shè)計(jì). 23 3.1引言

30、. 23 3.2 MIMO OFDM系統(tǒng)中信道時(shí)域響應(yīng)的LS估計(jì)及估計(jì)均方誤差. 25 3.3導(dǎo)頻非均勻分布時(shí)最優(yōu)導(dǎo)頻序列的設(shè)計(jì). 27 3.4仿真結(jié)果及討論. 34 3.5本章小結(jié). 36第四章OFDM系統(tǒng)中基于參數(shù)化模型的最大似然信道估計(jì). 37 4. 1引言. 37 4.2基于參數(shù)化信道模型下的OFDM系統(tǒng)等效基帶模型.38 4.3信道路徑數(shù)的檢測和路徑時(shí)延的初始化估計(jì). 39 4.4信道路徑時(shí)延與路徑增益的最大似然估計(jì). 41 4.4.1利用SAGE算法分解復(fù)雜的多參數(shù)最大似然估計(jì)問題. 42 4.4.2求解一個(gè)低復(fù)雜度的兩參數(shù)最大似然估計(jì)問題. 43 4.5算法復(fù)雜度分析. 44 4

31、.6仿真結(jié)果及討論. 45 4.7本章小結(jié). 48第五章OFDM系統(tǒng)中基于低通插值的時(shí)變信道估計(jì). 49 5.1引言. 49 5.2時(shí)變環(huán)境下OFDM系統(tǒng)的等效基帶模型. 50東南大學(xué)博士學(xué)位論文5.3基于低通插值的時(shí)變信道估計(jì). 515. 3. 1 P個(gè)選定時(shí)刻上的信道向量估計(jì).5. 3.2剩余時(shí)刻上基于低通插值的信道向量估計(jì) 5.4仿真結(jié)果及討論. 5.5本章小結(jié).第六章MIMO OFDM系統(tǒng)中基于分段線性模型的時(shí)變信道估計(jì)及相應(yīng)導(dǎo)頻序列的設(shè)計(jì). 6.1引言. 6.2時(shí)變環(huán)境下MIMO OFDM系統(tǒng)的等效基帶模型. 6.3基于信道分段線性假設(shè)下的MIMO OFDM系統(tǒng)模型. 6.4第一種時(shí)

32、變信道估計(jì)算法及相應(yīng)導(dǎo)頻序列的設(shè)計(jì). 6.4.1時(shí)變信道估計(jì)算法····························. 6.4.2導(dǎo)頻序列的設(shè)計(jì)··············&#

33、183;·················· . 6.5第二種時(shí)變信道估計(jì)算法及相應(yīng)導(dǎo)頻序列的設(shè)計(jì). 6.5.1信道估計(jì)算法. 6. 5.2導(dǎo)頻分布的設(shè)計(jì). 6.5.3導(dǎo)頻取值的設(shè)計(jì). 6.6仿真結(jié)果及討論. 6.6.1線性模型與Jakes模型的匹配性. 6.6.2導(dǎo)頻分布對估計(jì)算法的影響. 6.6. 3導(dǎo)頻取值對估計(jì)算法的影響. 6. 6. 4 MIMO系統(tǒng)中兩種所提估計(jì)算法的性能比較. 6.7本章小結(jié).第七章OF

34、DM系統(tǒng)中基于基擴(kuò)展模型的快時(shí)變信道估計(jì)及相應(yīng)最優(yōu)導(dǎo)頻序列的設(shè)計(jì). 7.1引言. 7.2基擴(kuò)展信道模型. 7.3基于基擴(kuò)展模型下的OFDM系統(tǒng)模型. 7.4快時(shí)變信道估計(jì)算法及相應(yīng)的最優(yōu)導(dǎo)頻序列設(shè)計(jì). 741決時(shí)變信道估計(jì)算法·····················. 7.4.2最優(yōu)導(dǎo)頻序列的設(shè)計(jì). 7.5仿真結(jié)果及討論. 7.6本章小結(jié).第八章結(jié)論. 8.1已取得的研究成果. 8.2

35、今后的研究工作.致謝.參考文獻(xiàn).作者攻讀博士學(xué)位期間發(fā)表和錄用的論文.符號定義符號定義矩陣或矢量的共扼轉(zhuǎn)置矩陣或矢量的轉(zhuǎn)置矩陣或矢量的共扼二L二E-)tr(A)一二一·(A)一(A)卞INONXMdiag Iv甲JAx mod yRef-IZ不小于x的最小整數(shù)不大于x的最大整數(shù)數(shù)學(xué)期望矩陣A的跡標(biāo)量x的絕對值歐幾里德范數(shù)矩陣A的逆矩陣A的偽逆NxN維單位矩陣NxM維的全零矩陣以向量v為對角線元素的對角矩陣取出矩陣A的對角元素使其成為一個(gè)列向量表示取余運(yùn)算，也即求解二一ny，其中。=二/川表示取實(shí)部整數(shù)集合東南大學(xué)博士學(xué)位論文縮略詞ADSLAsymmetric Digital Subs

36、criber Lines異步數(shù)據(jù)用戶鏈路AMPSAdvanced Mobile phone services高級移動電話服務(wù)BERBit Error Rate誤比特率CDMACode Division Multiple Access碼分復(fù)用CPCSICyclic Prefix循環(huán)前綴Channel State Information信道狀態(tài)信息DAB/DVBDigital Audio Broadcasting/ DigitalVideo BroadcastingDigital Advanced Mobile phone services數(shù)字音頻/視頻廣播D-AMPS數(shù)字高級移動電話服務(wù)DFTD

37、iscrete Fourier Transform離散傅立葉變換FIRFinite Impulse Response有限沖擊響應(yīng)GSMGlobal system for mobile communication全球移動通信系統(tǒng)HIPERLAN/2 HIgh PERformance LAN type 2ICIIntercarrier Interference載波間干擾IDFTInverse Discrete Fourier Transform離散傅立葉反變換Intersymbol InterferenceLeast Square碼間干擾最小二乘IslLsLMMSELinear Minimum M

38、ean Square Error線性最小均方誤差MIMOMultiple Input Multiple Output多輸入多輸出MLMaximum Likelihood最大似然MMSEMinimum Mean Square Error最小均方誤差MSEMean Square Error均方誤差VTTT縮略詞OFDMOrthogonal Frequency Division Multiplexing正交頻分復(fù)用SAGESpace-Alternating Generalized Expectation一廣義交互空間期望最大化MaximizationSISOSingle Input Single Output單輸入單輸出SNRSignal-to-Noise Ratio信噪比TDMATime Division Multiple Access時(shí)分復(fù)用WERWord Error Rate誤碼率WLANWireless Local Access Network無線局域網(wǎng)WSSUS