基于结构化稀疏特性的水声信道估计技术研究

发布时间：2018-06-02 00:48

本文选题：水声信道估计 + 正交频分复用　；参考：《江苏科技大学》2017年硕士论文

【摘要】：水声信道具有时-空-频变性、带宽有限、多径干扰严重,是迄今为止最复杂的无线信道。正交频分复用(OFDM)技术特别适合于带宽有限的水声信道,可有效对抗多径造成的码间干扰。随着通信距离的增加,信道状态不断恶化,造成接收信号在幅度、相位和频率上的失真。所以需要利用信道估计技术对信道状态进行跟踪,并对接收到的信号进行补偿。本文主要研究基于OFDM系统的水声信道估计技术。在传统的水声信道估计中,均假设是丰富多径信道,使用奈奎斯特采样定理对信道冲击响应进行采样,需要插入大量的导频进行信道估计,降低系统信息传输速率和频谱利用率。压缩感知理论(Compressive Sensing,CS)打破了传统的奈奎斯特采样定理,在采样的同时对信号进行压缩,使用少量的采样值就能高概率的重构出原始信号。使用标准CS理论进行估计是基于稀疏水声信道模型。然而大量海洋实验表明:由于海底介质的非均匀性,导致声线以块的形式进行传播,水声信道呈现出块结构稀疏特性。由于水声信道具有内在块结构稀疏特性,因此可以将信道估计问题转化为块结构稀疏信号重构问题,使用结构化压缩感知理论对信道进行块稀疏采样和重构。大幅降低了信道重构所需插入的导频数量,提高了系统频带利用率,且避免了对无用零抽头的估计,提高了重构的效率。本文首先介绍水声信道自身的物理特性和OFDM系统的基本通信原理,在射线理论的基础上构建水声相干多径信道模型,简单介绍了传统的最小二乘(Least Square,LS)信道估计方法。接着重点介绍压缩感知理论,在此基础上描述了块结构稀疏信号的概念及其相关的重构算法,重点介绍了经典的块正交匹配追踪(Block Orthogonal Matching Pursuit,BOMP)算法。引入多项正交匹配思想,将改进的BOMP算法应用于水声信道估计中。基于导频的信道估计方法计算简单,易于实现,得到了广泛的应用。本文使用基于导频的信道估计方法,将传统的LS,基于稀疏模型的正交匹配追踪(Orthogonal Matching Pursuit,OMP),基于块稀疏模型的BOMP及其改进的算法都应用到水声中,进行水声信道的重构。本文对水声信道估计进行系统的实验仿真。仿真结果表明:基于块结构稀疏模型的BOMP算法估计性能要优于传统的LS和OMP算法,插入少量的导频就能获得更优的估计性能,同时降低信道重构所需的时间;改进的BOMP算法在保证和BOMP算法重构精度一致的基础上,进一步降低重构所需的时间。
[Abstract]:Underwater acoustic signal props are sometimes space - frequency denaturation, limited bandwidth and serious multipath interference. It is the most complex wireless channel so far. Orthogonal frequency division multiplexing (OFDM) technology is especially suitable for the underwater acoustic channel with limited bandwidth. It can effectively combat intersymbol interference caused by multipath. With the increase of communication distance, the channel state is deteriorating, resulting in receiving signal in the channel. Distortion in amplitude, phase and frequency. Therefore, channel estimation techniques need to be used to track the channel state and compensate the received signals. This paper mainly studies the underwater acoustic channel estimation based on OFDM system. In the traditional underwater acoustic channel estimation, it is assumed that the multipath channel is rich rich, and the Nyquist sampling theorem is used for the letter. Compressive Sensing (CS) breaks the traditional Nyquist sampling theorem, compresses the signal at the same time of sampling, and uses a small amount of sampling values to reconstruct high probability. The original signal. The standard CS theory is based on the sparse underwater acoustic channel model. However, a large number of ocean experiments show that the acoustic line is propagated in block form due to the heterogeneity of the submarine medium, and the underwater acoustic channel presents a sparsity of block structure. The channel estimation problem is transformed into a block structure sparse signal reconstruction problem. The structured compressed sensing theory is used for the sparse sampling and reconstruction of the channel block. The number of pilots needed to be inserted in the channel reconstruction is greatly reduced, the utilization rate of the system is improved, and the estimation of the useless zero taps is avoided, and the efficiency of the reconstruction is improved. This paper first introduces the efficiency of the reconstruction. The physical characteristics of Shaoxing sound channel and the basic communication principle of OFDM system are used to construct the multipath channel model of underwater acoustic coherence on the basis of ray theory. The traditional Least Square (LS) channel estimation method is briefly introduced. Then the compression perception theory is introduced. On this basis, the concept of block structure sparse signal is described. And related reconstruction algorithms, the classic block orthogonal matching tracking (Block Orthogonal Matching Pursuit, BOMP) algorithm is introduced. The improved BOMP algorithm is applied to the underwater acoustic channel estimation by introducing a number of orthogonal matching ideas. The channel estimation method based on pilot is simple and easy to implement, and has been widely used. This paper makes a wide application. Using the channel estimation method based on pilot, the traditional LS, Orthogonal Matching Pursuit (OMP) based on sparse model, BOMP based on block sparse model and its improved algorithm are applied to underwater acoustic channel reconstruction. The experimental simulation of underwater acoustic channel estimation is carried out in this paper. The simulation results show that: The estimation performance of BOMP algorithm based on block structure sparse model is better than the traditional LS and OMP algorithm. Inserting a small number of pilots can obtain better estimation performance and reduce the time needed for channel reconstruction. The improved BOMP algorithm further reduces the time needed for reconstruction on the basis of ensuring the consistency of the reconstruction precision of the BOMP algorithm.
【学位授予单位】：江苏科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.3

【相似文献】