基于压缩感知的OFDM系统双选信道估计研究

发布时间：2018-01-15 03:19

本文关键词：基于压缩感知的OFDM系统双选信道估计研究　出处：《天津理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：高速移动下正交频分复用系统无线通信信道要经历时间和频率的双选择性衰落,为了使发送的数据经过双选信道后在接收端被正确地接收,必须对信道状态信息进行估计。由于双选信道在时延-多普勒域呈现出稀疏性,压缩感知技术得以应用到双选信道稀疏信道估计中。本文主要研究了基于压缩感知的双选信道估计。主要研究内容包括:(1)提出了双选信道估计下最优导频模式的选取方法。OFDM系统采用压缩感知技术进行双选信道估计时,随机选取的导频模式易导致压缩感知的观测矩阵重构能力不稳定,影响信道估计性能。可通过多种相关性标准来描述双选信道的观测矩阵,根据观测矩阵的相关性最小准则选取最优导频模式可改善信道估计性能。仿真结果表明比起随机放置导频和等距放置导频,所选最优导频模式的信道估计性能更好。(2)基于OMP、ROMP算法提出了两种改进恢复算法。当稀疏度提高时,压缩感知中ROMP算法的复杂度增大,为了降低复杂度,提出了两种改进算法。一种是有严格计算约束的改进恢复算法,该算法每次迭代选择固定数目的原子使支撑集为非奇异矩阵来降低原子选择和LS计算上的复杂度,并且每次迭代更新支撑集来保证精度;另一种改进恢复算法利用QR矩阵分解来避免ROMP中LS中的伪逆运算,降低了计算复杂度。(3)ICI存在的情况下提出了基于压缩感知技术的双选信道估计方法。ICI存在时,双选信道系数矩阵的主对角线及其他对角线皆为非零值,而传统信道估计方法忽略了主对角线之外的其他部分。根据信道系数矩阵的每条对角线在时延-多普勒域都具有稀疏性,利用压缩感知理论,可对信道系数矩阵的所有对角线进行估计,并通过选取适当的信道系数矩阵有效带宽的大小,获得最优的估计性能。仿真结果表明该估计方法可提高ICI存在下双选信道的估计性能。
[Abstract]:High speed mobile selective wireless orthogonal frequency division multiplexing system to wireless communication channel through time and frequency fading, in order to send the data after the doubly selective channel can be received correctly by the channel state information must be estimated. Because in the doubly selective channel delay Doppler domain shows a sparse, compressed sensing technology applied to the doubly selective channel sparse channel estimation. This paper mainly studies the doubly selective channel estimation method based on compressed sensing. The main research contents include: (1) put forward the double channel estimation method for.OFDM system under the selected optimal pilot mode using double channel estimation of compressed sensing technology, pilot pattern randomly selected easily lead to capacity matrix observation of reconstruction of compressed sensing instability, affecting the performance of channel estimation. The observation matrix can be described through a variety of doubly selective channel correlation standard, according to the observation The criterion of minimum correlation matrix to select the optimal pilot pattern can improve the channel estimation performance. The simulation results show that compared with randomly placed pilot and placed equidistant pilot channel estimation performance better, the selected optimal pilot mode. (2) based on OMP ROMP algorithm, puts forward two kinds of complex algorithm. When the sparsity increases. ROMP, compressed sensing algorithm complexity increases, in order to reduce the complexity, two improved algorithms are proposed. One is the recovery algorithm to improve the rigorous calculation of constraints, the algorithm in each iteration to choose a fixed number of atoms of the support set to reduce the complexity of atom selection and LS calculation of the non singular matrix. And each iteration to ensure the accuracy of the support set; another improved restoration algorithm using QR matrix decomposition to avoid ROMP LS pseudo inverse operation, reduces the computational complexity. (3) the presence of ICI is proposed based on the pressure Doubly selective channel estimation method for.ICI compressed sensing technology exists, the main diagonal coefficient matrix of the doubly selective channel and other diagonal is nonzero, and ignore the other part of the main diagonal method outside the traditional channel estimation. According to each diagonal coefficient matrix of the channel sparsity is in delay Doppler domain, using the compressed sensing theory, can all of the diagonal coefficient matrix of channel estimation, and by selecting the effective bandwidth of channel coefficient matrix of appropriate size, to obtain the best estimation performance. The simulation results show that the estimation performance of the estimation method can improve the presence of ICI doubly selective channel.

【学位授予单位】：天津理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.53

【相似文献】