MIMO-OFDM系统的联合频偏与信道估计技术研究

发布时间：2019-06-20 04:03

【摘要】：正交频分复用(Orthogonal Frequency Division Multiplexing, OFDM)技术能够有效地对抗多径衰落,同时具有较高的频谱效率,因此是宽带移动通信系统最佳的调制和多址接入候选方案。多输入多输出(Multiple-Input Multiple-Output, MIMO)技术能够有效地利用多径来提高系统容量(复用增益)或链路可靠性(分集增益),是极大提高通信系统频谱效率的有效手段。而两者联合而成的MIMO-OFDM技术,集两者之长于一身,已成为时下热门移动通信系统的标准配置。作为多载波系统,OFDM对载波频率偏移(Carrier Frequency Offset, CFO)非常敏感。由多普勒频移和收发端晶振不匹配引起的CFO,会破坏子载波间的正交性,造成子载波间干扰,导致系统性能严重恶化。同时,为了实现相干解调,准确的信道冲激响应(Channel Impulse Response, CIR)信息也是不可或缺的。在通信系统中,这两个未知参数经常是同时存在并耦合在一起的,因此研究联合CFO和CIR估计算法具有现实意义。针对时不变频选衰落信道MIMO-OFDM系统,论文在前人研究的基础之上,提出了一种“多天线、多径”模型；基于此模型,推导了最大似然联合估计和相应的CRLB限：接着为了降低估计的复杂度,论文研究了基于期望最大化型算法(Expectation-Maximization,EM)的迭代联合估计,仿真的结果表明EM型算法能够有效地逼近克拉美罗下限(Cramer-Rao Lower Bound, CRLB);最后,论文针对稀疏信道环境,采用子空间匹配(Subspace Pursuit, SP)压缩感知算法,修正了信道估计部分的方法,仿真的结果表明修正后的迭代联合估计具有更好的估计性能和更低的迭代复杂度。针对时变频选衰落信道MIMO-OFDM系统,论文分析了时变信道的统计特性,给出了时变信道表示方法——基扩展模型(Basis Expansion Model, BEM),并仿真比较了几种常见BEM的表示性能；然后在前人研究的基础之上,论文先研究了SISO-OFDM系统的系统模型,并分别给出了频域和时域两种联合估计方法；最后针对MIMO-OFDM系统,论文提出了一种“多天线、多径、时变”模型,从前面的时域联合估计方法中拓展出基于最大似然准则和最大后验准则的两种估计方法,仿真结果表明基于时变相关特性的最大后验方法具有更好的估计性能。最后,简单梳理了本文的研究思路和取得的成果,并针对不足之处,探讨了以后的研究方向和改进思路。
[Abstract]:Orthogonal Frequency Division Multiplexing (Orthogonal Frequency Division Multiplexing, OFDM) technology can effectively combat multi-path fading and has high spectral efficiency, so it is the best modulation and multiple access candidate scheme for broadband mobile communication systems. Multi-input multiple-output (Multiple-Input Multiple-Output, MIMO) technology can effectively use multi-path to improve system capacity (reuse gain) or link reliability (diversity gain). It is an effective means to greatly improve the spectral efficiency of communication systems. The MIMO-OFDM technology, which combines the two technologies, has become the standard configuration of popular mobile communication systems. As a multi-carrier system, OFDM is very sensitive to carrier frequency offset (Carrier Frequency Offset, CFO). The CFO, caused by Doppler frequency shift and crystal oscillator mismatch at the transceiver will destroy the orthogonality between subcarriers and cause inter-carrier interference, resulting in a serious deterioration of system performance. At the same time, in order to realize coherent demodulation, accurate channel impulse response (Channel Impulse Response, CIR) information is also indispensable. In communication systems, these two unknown parameters often exist at the same time and are coupled together, so it is of practical significance to study the joint CFO and CIR estimation algorithms. Based on the previous research, a "multi-antenna, multi-path" model for time-invariant frequency selective fading channel MIMO-OFDM system is proposed in this paper. Based on this model, the maximum likelihood joint estimation and the corresponding CRLB limit are derived. Then, in order to reduce the complexity of the estimation, the iterative joint estimation based on the expected maximization algorithm (Expectation-Maximization,EM) is studied. The simulation results show that the EMtype algorithm can effectively approach the Cramer-Rao Lower Bound, CRLB); lower limit (Cramer-Rao Lower Bound, CRLB);). Finally, the subspace matching (Subspace Pursuit, SP) compression sensing algorithm is used to modify the channel estimation method for sparse channel environment. The simulation results show that the modified iterative joint estimation has better estimation performance and lower iterative complexity. For time-varying frequency-selective fading channel MIMO-OFDM systems, the statistical characteristics of time-varying channels are analyzed, and the base spread model (Basis Expansion Model, BEM), which is a time-varying channel representation method, is given and the representation performance of several common BEM is simulated and compared. Then, on the basis of previous studies, the system model of SISO-OFDM system is studied, and two joint estimation methods in frequency domain and time domain are given respectively. Finally, for MIMO-OFDM system, a "multi-antenna, multi-path, time-varying" model is proposed, and two estimation methods based on maximum likelihood criterion and maximum posterior criterion are extended from the previous joint time-domain estimation methods. The simulation results show that the maximum posterior method based on time-varying correlation characteristics has better estimation performance. Finally, the research ideas and achievements of this paper are briefly combed, and the future research direction and improvement ideas are discussed in view of the shortcomings.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.53

【共引文献】