求博士论文《OFDM无线通信系统中的定时恢复和信道估计算法》

我也要看看：P
truecafan@gmail.com

他是哪个试验室的，直接问他的师弟要不就好了。

曹志刚老师的学生。但是我不认识他师弟啥。呵呵。你找到了，给我也发一份。谢谢！网上居然没有。很奇怪。

谢谢，同求
cqupt.yale@gmail.com

事实上，师弟也没有啊
.10

我问他的师妹要了，师妹总有吧。

要了之后，请一定记得给我发一份。谢谢！再次写email:  hustwang@gmail.com

同求，谢谢
  team2005@sohu.com

他师妹不理我也没办法。。。

可以直接到清华图书馆办理馆际

呵呵，只找到一个摘要
杨宝国


论文题目：OFDM无线通信系统中的定时恢复和信道估计算法

作者简介：杨宝国，男，1973年01月出生，1996年07月师从于清华大学曹志刚教授，于2001年01月获博士学位。


摘 要
下一代移动无线通信系统的目标是实现无所不在的、高质量的、高速率的移动多媒体传输。但是为了实现这一目标，面临许多技术挑战。例如，移动无线通信系统面临的是十分恶劣的无线信道。稳健的移动无线通信系统不仅需要克服大的路径损耗，以及非常严重信号衰落，还要克服由于大的多径时延扩展而引起的符号间干扰。而正交频分复用(Orthogonal Frequency Division Multiplexing, OFDM)技术则是一种很有前途的、可克服信道时延扩展的传输手段。但是，OFDM系统对于同步误差和信道估计误差都很敏感。针对上述问题，本文重点研究了OFDM接收机中的定时恢复和信道估计算法。分别提出了一种包括码元同步和采样时钟同步在内的OFDM系统定时恢复方案，一种适用于稀疏多径衰落信道、基于参数化信道模型的信道估计算法，以及一种低复杂度、基于加窗离散傅立叶变换的、针对非参数化信道模型的最小均方误差的信道估计算法。
在第二章中，我们首先简要介绍有关无线通信信道的基本概念，并主要讨论多径信道的时延扩展对接收信号的影响。
第三章主要讨论OFDM系统接收机的设计问题。首先，我们介绍OFDM的基本原理，包括其理想的信号模型。特别是，我们引入了OFDM的块传输模型，从中我们可以看到循环前缀对于OFDM系统接收机的简化起着关键作用。然后，我们讨论当考虑了各种非同步因素后，OFDM系统的实际传输模型。其中，我们将OFDM接收机分成内接收机和外接收机来讨论。接着，我们分析了各种非理想传输条件对OFDM接收机性能的影响。最后，我们总结了实现OFDM内接收机中各部分功能的主要算法，并给出了一个OFDM内接收机的总体设计。
论文第四章提出了一种包括实现OFDM码元同步和采样时钟同步的定时恢复算法。在OFDM传输系统中，同步的任务包括载波同步和定时恢复。而定时恢复又可进一步划分为OFDM码元同步和采样时钟同步。码元同步的目的就是找到正确的FFT窗位置。OFDM码元的循环前缀可被用来做码元同步。但是在多径衰落信道中，由多径信道引起的ISI已破坏了循环前缀的重复特性。在这种情况下，基于循环前缀的码元同步方法的性能将得不到保证。如果码元定时误差超出一定范围后，不准确的码元同步将引入ISI，破坏各子载波之间的正交性，使OFDM系统性能下降。更严重的是，常用在相干OFDM系统中，基于Wiener滤波器的信道估计算法对码元同步误差非常敏感。所以，相干OFDM系统对码元同步准确性的要求更高。另一方面，采样时钟同步的主要目的是使接收机和发射机的采样时钟频率保持一致。采样时钟频率偏差将导致ICI。采样时钟频率偏差还将导致码元定时的漂移，进一步恶化码元同步的问题。而在已往算法中，采样时钟同步和码元同步没有有机地结合在一起。所以，我们提出一种基于导频的，包括码元和采时钟同步的OFDM传输系统定时恢复方法。在该方法中，我们用路径时延估计来提高基于相关算法的码元定时估计的准确性，并用一个反馈环路来跟踪采样时钟和锁定码元定时。我们提出了两种非相干反馈环路：DLL和MLL。该两种方法都可看成是在AWGN信道下，对码元定时和载波相位进行联合最大似然估计的递归解。我们分析了在AWGN信道下两种环路的跟踪误差均方值，并推导了对码元定时估计方差的CRB界。分析显示，MLL的跟踪误差的均方值渐近于该CRB界。而可以看成近似解的DLL，其性能相对于非近似解的MLL下降很小。另外，我们分别对以上算法在AWGN信道、非频率选择性衰落信道和多径衰落信道中进行了仿真，来验证其性能。结果表明，相对于传统的基于相关算法的码元同步方法而言，本章中提出的方法可将定时估计误差的均方值减小几个数量级。而且，这里提出的反馈环路方法可以跟踪由于采样时钟频差引起的码元定时漂移。
论文第五章提出一种适用于稀疏多径衰落信道改进的OFDM系统信道估计算法。在无线通信中，我们通常是用多径传播来描述信道的。例如，在一个宏区中，当基站天线位置较高时，这时的多径信道将主要由2到6条反射路径构成。我们可以使用参数化的信道模型来描述这种信道。在该模型中，各路径只包含时延和复增益系数两个参数。这种参数化的方法可使信道估计问题的维数明显减少。而我们知道，当训练数据数量相同时，减小所要估计问题的维数便可提高估计的准确性。特别是本章显示，对于稀疏多径衰落信道，当基于参数化信道模型来构造信道相关矩阵时，可以明显减少信道相关矩阵的信号子空间维数。对于MMSE信道估计器，减小信号子空间的维数可直接提高估计器的性能。另外在移动环境中，路径时延变化是很慢的，而路径增益系数的幅度和相位变化都非常快，一般认为它们是按Rayleigh衰落的。这一特性可进一步简化信道估计器的设计。所以，我们提出使用一种利用导频并基于参数化信道模型的信道估计方法。基于该模型的信道估计器就是要估计包括路径数、各路径的时延及复数增益系数在内的这些信道参数。首先，我们采用最小描述长度(Minimum Description Length, MDL)准则来检测信道中的路径数。然后，利用估计信号参数的旋转不变法(Estimation of Signal Parameters by Rotational Invariance Techniques，ESPRIT)来对各路径时延进行初始估计。因为路径时延的慢变化，所以我们提出使用路径间干扰抵消(Inter-Path Interference Cancellation, IPIC)的延迟锁定环路(DLL)来跟踪信道路径时延。最后，利用信道的路径时延信息，我们推导出一个对信道频域响应的MMSE估计器。仿真结果显示MDL准则和ESPRIT方法可自适应地对信道参数进行初始估计。而IPIC DLL也被证明是一种估计和跟踪路径时延的有效方法。分析和仿真还表明，这里提出的基于参数化信道模型的信道估计算法相对于基于非参数信道模型的方法，可明显改善OFDM传输系统在稀疏多径信道下的性能。
在论文第六章中，我们针对非参数化信道模型，提出并分析了低复杂度的基于加窗DFT的MMSE信道估计算法。在非参数化信道模型方法中，最小均方误差(Minimum Mean Square Error, MMSE)准则经常被用来设计OFDM系统的信道估计器，但是理想的MMSE估计器的复杂度一般都很高，而利用离散傅立叶变换(Discrete Fourier Transform, DFT)来进行OFDM信道估计可大大降低估计器的复杂度。基于DFT的方法既可用到内插的情形中，又可用到非内插的情形中。用于内插时，DFT是一种简单的，并且计算上十分高效的内插算法。但是完全精确的内插需要信道中各路径时延必须按OFDM采样间隔来分布。实际中，信道的多径时延一般不是按OFDM采样间隔分布的，甚至信道的功率时延分布根本不是离散的，而是连续的。这种情况下，如果在进行基于DFT的内插之前，我们不用一个窗函数对观测到的信道频率响应向量进行加窗处理，则内插后，由能量泄漏而引起的混叠现象将导致信道估计的错误平底(Error floor)。另外，由于导频信号通常对信道频率响应进行过采样，所以在接收端得到的等效信道冲击响应向量中，信道能量主要集中在一个较小的范围内，而噪声能量则分布在整个向量范围内。所以，我们可以在时域对等效信道冲击响应向量使用一个加权函数来减小信道噪声的影响。对于非内插情形，基于DFT的方法也是利用了DFT实现的低复杂度和信道能量在等效信道冲击响应向量中分布相对集中的特性。在本章中，我们针对内插和非内插不同的情形，提出一种低复杂度的，基于加窗DFT的MMSE信道估计算法。首先，在频域，我们使用一个推广的Hanning窗函数，对由导频得到的信道频率响应观测向量进行加窗操作，来减小能量泄漏。另外，在时域，我们还对得到的等效信道冲击响应向量使用一个加权函数来减小信道噪声的影响。对于一个给定的频域窗函数，我们是通过使信道估计均方误差(Mean Square Error, MSE)最小，来选取时域加权函数的。另外，由于信道估计的MSE还依赖于频域窗函数的形状，我们将通过搜索来找到使信道估计MSE达到最小的最佳窗函数形状。分析和仿真表明，采用频域加窗可消除内插情形中信道估计的错误平底，并可在非内插情形中带来更好的噪声滤除性能。而在时域中的MMSE加权是一种抑制信道噪声和改进信道估计性能的很有效的方法。而且，分析和仿真结果还显示，建议方法的性能接近于最优的MMSE信道估计器的性能，但复杂度却减小很多，这是因为在建议的方法中，我们使用了快速算法IFFT/FFT来实现IDFT/DFT，并且在频域的加窗过程和在时域的MMSE加权过程，都是简单的单个元素与单个元素相乘的操作。
第七章是论文的总结和结论，并对进一步工作提出了建议。

Timing Recovery &Channel Estimation Algorithms for
Wireless Communication Systems Using OFDM
Abstract
The next-generation wireless personal communication systems are expected to provide ubiquitous, high-quality, and high-rate mobile multimedia transmission. However, to achieve this objective various technical challenges must be overcome. For example, the deployment of broadband wireless access systems would require a transmission technique which can mitigate the detrimental effects of the inter-symbol interference (ISI) caused by the multipath fading channels. In recent years, there has been a lot of interest in applying Orthogonal Frequency Division Multiplexing (OFDM) in wireless systems because of its various advantages in lessening the severe effects of ISI. However, the OFDM system is vulnerable to synchronization errors and channel estimation errors. How to effectively do the synchronization and do the channel estimation at the OFDM receiver are the important issues, which need be addressed in OFDM systems. Hence, the dissertation focuses on the timing recovery and channel estimation algorithms design for OFDM receivers.
Chapter 2 gives a brief introduction about the characteristics of the wireless communication channels, and mainly discusses the inter-symbol interference caused by the channel multipath delay spread.
In Chapter 3, we discuss the OFDM receiver design issues. We first introduce the basic principles about OFDM transmission, including the ideal signal model. Specifically, we use the block transmission model to represent the OFDM system. After that, we consider several unsynchronized factors impacts on the OFDM transmission and introduce a real OFDM transmission model. We then divide the OFDM receiver as inner-receiver and outer-receiver. Finally, we present an inner-receiver design of OFDM system.
In Chapter 4, we propose a scheme for performing timing recovery that includes symbol synchronization and sampling clock synchronization in OFDM systems. In OFDM transmission systems, the synchronization tasks include carrier frequency synchronization and timing recovery that can be further divided into symbol synchronization and sampling clock synchronization. The purpose of symbol synchronization is to find the correct position of the fast Fourier transform (FFT) window. Symbol synchronization may be done at the receiver with the aid of the dedicated training symbols. The cyclic property of the guard interval preceding the OFDM symbol can be also evaluated for symbol synchronization, thus reducing the need for training symbols. In multipath fading channels, however, the guard interval is corrupted by ISI and the periodic property is destroyed. Consequently, correct symbol synchronization cannot be guaranteed in the case of ISI. If the symbol timing error is outside of the ISI free range in the guard interval, the inaccurate symbol timing can cause ISI that destroys the orthogonality of the sub-carriers and degrades the performance of OFDM systems. In addition, the performance of the channel estimation via interpolation, commonly used for coherent OFDM systems, can be essentially degraded by the symbol timing errors. Hence, more accurate symbol timing synchronization methods are needed to fulfill the synchronization requirement in coherent OFDM systems. In contrast to the symbol synchronization case, the purpose of sampling clock synchronization is to align the receiver sampling clock frequency to that of the transmitter. The sampling clock frequency error can cause ICI. Moreover, the sampling clock frequency error can result in a drift in the symbol timing and can further worsen the symbol synchronization problems. Thus, sampling clock synchronization is also an important issue that needs to be addressed in OFDM systems. In current methods, symbol synchronization and sampling clock synchronization are dealt with separately. In this Chapter, we propose a timing recovery scheme based on pilot sub-carriers for OFDM systems, which can solve the symbol synchronization and sampling clock synchronization issues simultaneously. As pilot sub-carriers are used in most coherent OFDM systems for synchronization and channel estimation purposes, our scheme can be implemented without additional overhead for these systems. In this scheme, we use the correlation method based on the guard interval to do the coarse symbol synchronization. A path time delay estimation method is then employed to further improve the accuracy of the coarse symbol synchronization. Finally we use a delay-locked loop (DLL) to do the sampling clock synchronization and to maintain the symbol timing. Other feedback loop MLL is also proposed. We derive the both techniques from the joint maximum-likelihood (ML) estimation of the symbol timing and carrier phase in the AWGN channel. We apply the derived algorithm to both AWGN and various multipath fading channels and study their performance via simulation. Analysis shows that the tracking error of MLL approaches to the CBD bound. Even though the DLL is not optimal for the multipath fading channels, by employing this scheme, the mean square values of the symbol timing estimation error can be reduced by several orders of magnitude compared with the traditional correlation methods. In addition, the proposed scheme is capable of tracking the drift in the symbol timing caused by the sampling clock frequency offsets.
In Chapter 5, we present a novel channel estimation algorithm for OFDM mobile communication systems using pilot sub-carriers. This algorithm is based on a parametric channel model. In OFDM systems, the channel estimation can be achieved by exploiting the correlation of the channel frequency response at different frequencies and times. The channel estimators for OFDM systems have been proposed based on frequency domain filtering and time domain filtering. These methods do not make any assumptions about the channel model and hence the dimension of the estimation problem can be quite large. However, the radio channel in a wireless communication system is often characterized by the multipath propagation. In large cells with high base station antenna platforms, the multipath propagation is aptly modeled by a few dominant specular paths, typically two to six. Moreover, the high-speed data transmission in wireless communications potentially results in a sparse multipath fading channel. The sparsity of a multipath channel can be defined as the ratio of the time duration (in OFDM samples) spanned by the multipaths to the number of the multipaths. A parametric channel model can then be used to represent this type of channels. When the channel correlation matrix is constructed based on the parametric channel model, the signal subspace dimension of the correlation matrix can be effectively reduced. Accordingly, the channel estimator performance can be improved. The parametric channel model approach has been applied to the Global System for Mobile Communications (GSM) system and the high-speed digital video broadcast system to improve the channel equalizer and estimator performance. It should be also noted that in mobile communications the multipath time delays are slowly time-varying. In contrast, the amplitude and relative phase of each path are relatively fast time-varying and subject to (Rayleigh) fading. We can thus take this into account in designing the channel estimator. In this Chapter, we propose an improved channel estimation method for OFDM transmission over the sparse multipath fading channels using pilot sub-carriers. The channel estimator is based on a parametric channel model. That is, the channel frequency response of the multipath fading channel is modeled as the Fourier transform of a multipath finite impulse response. The channel estimator is derived to estimate the parameters which include the time delays, gains, and phases of the paths. Specifically, we first use the minimum description length (MDL) criterion to detect the number of paths in the channel. Then, we use the Estimation of Signal Parameters by Rotational Invariance Techniques (ESPRIT) to estimate the initial multipath time delays. Because of the slow time-varying nature of time delays, we propose an inter-path interference cancellation (IPIC) delay locked loop (DLL) to track the channel multipath time delays. With the multipath time delays information, a MMSE estimator is derived to estimate the channel frequency response. The simulation results show that the MDL criterion and the ESPRIT method can adaptively estimate the initial channel parameters. Further, the IPIC DLL is shown to be an effective way to estimate and track the multipath time delays. Analysis and simulation results also demonstrate that the proposed channel estimation algorithm gives a substantial performance improvement in MSE over the non-parametric channel model based methods for the sparse multipath fading channels.
In Chapter 6, low-complexity windowed discrete Fourier transform (DFT) based MMSE channel estimators are proposed and analyzed for both the interpolation and non-interpolation cases for OFDM mobile communications systems. The Minimum Mean Square Error (MMSE) interpolator and filter have been proposed to do channel estimation for OFDM systems. However, the complexity of the optimal MMSE estimator is usually high. The discrete Fourier transform (DFT) based channel estimator gives us an alternative and low-complexity choice. The DFT based method can be used for both the interpolation case and the non-interpolation case. For the interpolation case, the DFT is a simple and computationally efficient approach for performing interpolation. However, the perfect interpolation of a N-sample complex sequence by DFT not only requires that the sequence represents N equispaced samples of a continuous signal that is band-limited below the Nyquist limit, but also that the signal has a discrete spectral density distribution. In the application of the channel estimation for the OFDM systems, the above discrete spectral density requirement means that the channel multipath time delays must be sample-spaced. In practice, while the channel impulse response has a finite duration that is below the Nyquist sampling rate limit in the frequency domain, the channel multipath time delays will, in general, be non-sample-spaced or the channel will, in general, have a continuous rather than discrete power delay profile. In this case, if no data windowing is applied to the channel frequency response observation vector before the interpolation by DFT, the aliased spectral leakage can cause an error floor. In addition, since the channel frequency response is usually oversampled by the pilot sub-carriers, the channel power will be concentrated to a relatively small number of samples in the effective channel impulse response vector. On the other hand, the noise power is spread over the vector. Hence, a weighting function can be applied to the vector to suppress the channel noise. For the non-interpolation case, the DFT based estimators also employ the low-complexity property of DFT and the channel power concentration property of the effective channel impulse response. In this Chapter, we present the low-complexity windowed DFT based MMSE channel estimators for both the interpolation and non-interpolation cases. A generalized Hanning window is first applied to the channel frequency response observation vector to reduce the spectral leakage. Moreover, a weighting function is applied to the effective channel impulse response. The weighting function is chosen so that the MSE between the channel frequency response and its estimate is minimized for a given window function. Since the resulting MSE also depends on the window shape, the optimal generalized Hanning window shape is also searched to minimize the MSE. Analysis and simulation results show that the data windowing can eliminate the error floor for the interpolation case, and can bring about better noise filtering performance for the non-interpolation case. Moreover, the MMSE weighting is an effective technique to suppress the channel noise and improve the channel estimation performance. It is shown that the optimal MMSE weighting functions always exist for both cases. It has been also shown that the proposed method performance is close to the conventional optimal MMSE channel estimator and is much better than the direct DFT based estimator. Furthermore, it has much lower complexity than the optimal MMSE estimator because the IDFT/DFT transforms can be implemented with the fast algorithms IFFT/FFT and the MMSE weighting operation in the time domain is a simple element-by-element multiplication.
Finally, Chapter 7 summarizes the dissertation and gives some suggestions about future research directions about OFDM.

　

ESPRIT.....
我喜欢：）

正解
.10

怎么办?可以搞到电子版吗?

馆际?
他不是thu的吗？

她师妹说没有电子版地。。。。。