【作者】 王德刚；

【导师】 孙即祥； 魏急波；

【作者基本信息】 国防科学技术大学 ， 电子科学与技术， 2008， 博士

【摘要】 正交频分复用(OFDM)以及它和多输入多输出(MIMO)的结合—MIMO-OFDM是目前宽带无线通信非常具有潜力的传输技术之一,然而,OFDM和MIMO-OFDM技术对载波频偏(CFO)和相位噪声(PHN)非常敏感,CFO和PHN会破坏OFDM子载波的正交性,严重恶化系统性能。因此,有效对抗CFO和PHN成为OFDM和MIMO-OFDM技术实现的一个重要挑战。论文主要针对OFDM和MIMO-OFDM系统实现中的CFO估计和PHN抑制方法进行了研究。论文的主要工作和创新之处在于:1、针对准静态多径信道环境下的OFDM系统,提出了一种低复杂度的盲CFO估计算法,该算法利用解调数据二阶统计量矩阵的对角能量最大化准则得到CFO估计,有效提高了CFO估计精度。该算法经推导可得出估计式的闭环求解表达式,具有较低的实现复杂度。2、在时变多径信道环境下,提出了一种面向判决(DD)的联合CFO和信道盲估计算法,该算法基于空间选择性期望最大化(SAGE)算法思想,采用迭代的方式得到CFO和信道的精确估计,并通过近似得到闭环的CFO估计求解表达式,有效降低了算法复杂度。3、在准静态多径信道环境下,针对存在PHN的OFDM系统提出了一种面向判决的载波间干扰(ICI)迭代抑制算法,所提算法利用Huber滤波思想,改进了传统PHN最大似然(ML)迭代抑制方法的反馈函数,从而有效减弱了判决误差对迭代抑制性能的影响,提高了ICI系数估计精度,具有较好的PHN抑制性能。为降低ML准则下ICI抑制方法的复杂度,论文提出了一种基于优化最小化(MM)算法的迭代ICI抑制算法,该算法利用优化不等式理论得到ICI估计中ML代价函数的替代函数,降低了ML迭代估计中的矩阵求逆维数,有效降低了运算复杂度。4、在准静态多径信道坏境下,针对贝尔实验室分层空时(V-BLAST)复用结构的MIMO-OFDM系统模型,提出了一种基于子空间投影准则的盲CFO估计算法,该算法引入Hadamard积的概念,建立了接收信号子空间和基于离散傅立叶变换(DFT)的时频变换子空间结构,通过分析CFO对时频Hadamard积空间和信号Hadamard积空间关系的影响,得到了与CFO相关的子空间投影代价函数,并通过对投影范数的最大化得到了CFO的估计,估计性能优于已有的MIMO-OFDM盲CFO估计方法。5、对于OFDM和MIMO-OFDM系统,关于CFO的影响和直观的BER性能分析在现有文献中已有很详尽的论述,而PHN的影响和性能分析仍有待深入研究。论文重点分析了MIMO-OFDM系统在PHN影响下的性能,通过合理近似,给出了不同基带调制方式下BER性能的闭环表达公式,对于实际MIMO-OFDM系统设计实现中PHN抑制方法的选择提供了简单直观的依据。6、MIMO-OFDM系统的PHN抑制目前还只针对共同相位误差(CPE)进行抑制。由于忽略ICI的影响,CPE抑制对系统性能改善并不明显。为进一步提高PHN抑制性能,在V-BLAST结构的系统模型下,论文将OFDM系统基于ML的ICI抑制算法扩展到MIMO-OFDM系统。与单纯进行CPE补偿的方法相比,所提算法对MIMO-OFDM系统PHN抑制性能有较大提升。更多还原

【Abstract】 Orthogonal Frequency Division Multiplexing (OFDM) and multiple-input multiple-output OFDM (MIMO-OFDM) is a promising technique for broadband wireless communication currently. However, one of the main drawbacks of OFDM and MIMO-OFDM is that it is highly sensitive to carrier frequency offset (CFO) and phase noise (PHN). It is a significant challenge to suppress the effect of CFO and PHN in implementation of OFDM and MIMO-OFDM systems. This dissertation focuses on CFO estimation and PHN mitigation methods for implementation of OFDM and multiple-input multiple-output (MIMO) OFDM systems. The main contributions of the dissertation are as follows:1. A blind CFO estimation algorithm with low complexity was proposed for OFDM in quasi-static multipath channel. The likelihood function of CFO was extracted from signal covariance matrices in frequency domain, which could gave accurate CFO estimation. A closed-form solution could be manipulated with proposed algorithm, so it was computational efficient.2. A decision directed (DD) joint CFO and channel estimator was proposed for OFDM system in time-variant multipath channel. Based on space-alternating generalized expectation-maximization (EM) algorithm, CFO and channel state information were estimated iteratively. Furthermore, a close-form solution for CFO updating was derived by approximation.3. In quasi-static multipath channel, an iterative intercarrier interference (ICI) mitigation algorithm based on decision data was proposed for OFDM system with PHN. By using Huber filter, the conventional maximum likelihood feedback function in iteration was reformed, which could suppress the performance impairment due to decision error. So ICI estimation became more accurate and PHN mitigation has better performance. To reduce complexity in ICI mitigation based on ML criterion, a Majorization Minimize (MM) iterative algorithm for ICI mitigation was derived. By replacing ML cost function for ICI estimation with surrogate function, which is the majorization function of the ML cost function, the proposed PHN mitigation reduced orders of matrix to be inverted and has a lower complexity than the conventional ML PHN mitigation.4. Focus on V-BLAST space diversity multiplexing model in MIMO-OFDM systems, a blind CFO estimator was derived based on the subspace projection approach in quasi-static multipath channel. The proposed algorithm built sub-space of received signal and DFT based on Hadamard product concept. The projection cost function was obtained by analyzing the effect between CFO and Hadamard product sub-space of received signals and time-frequency domain transforming, then CFO was estimated accurately by maximization the Frobenius norm of projection. The proposed estimator has better performance than existing algorithms in literature.5. The performance analysis for OFDM and MIMO-OFDM systems with CFO were discussed detailedly in many literatures, but the effect and performance analysis of PHN were waiting for being studied deeply. The dissertation analyzed BER performance for MIMO-OFDM systems with PHN especially. Then, by reasonable approximate, a closed-form expression of BER was presented, which provided a simple and direct reference for choosing mitigation methods for implementation of MIMO-OFDM system.6. The PHN mitigation methods for MIMO-OFDM systems were focused on common phase error (CPE) mitigation up to now. CPE mitigation only for MIMO-OFDM systems could not give satisfying performance due to neglecting ICI component. To improving PHN mitigation performance further, the dissertation extended the ICI mitigation from OFDM systems to MIMO-OFDM systems based on V-BLAST system model. Compared with CPE mitigation method, the proposed mitigation algorithm can significantly improve the performance of PHN mitigation for MIMO-OFDM.更多还原