基于压缩感知的信道估计关键技术研究

发布时间：2018-01-20 01:39

本文关键词： 压缩感知信道估计稀疏性信号检测贪婪算法　出处：《电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：压缩感知是在本世纪所提出最重要的理论之一,它声称可以将稀疏信号从比奈奎斯特采样率得到地更少的样本点中恢复出来。压缩感知理论是在图像处理的背景下,发展出来的数据采集和压缩技术,革新了传统对信号的采集和存储的手段。因此它不仅在图像处理方面有着重要作用,在其他一些领域,如信道估计,信号采样以及数据压缩等,压缩感知理论都有极大的应用价值。本文将主要介绍压缩感知的基本理论以及压缩感知在信道估计中的应用。在无线通信中,在障碍物物较多的无线通信环境下,发送信号通常经过多条不同延时的路径到达接收者。多经无线信道通常被建模为一个线性滤波器,接收到的信号为发送信号经不同衰减和延迟的线性叠加。实测数据表明,传输路径的冲激响应在时域上可看作是一个近似稀疏的信号。因此,可以采用压缩感知技术对信道进行估计。由于考虑了信道的稀疏性,所以可以得到比传统信道估计方法更准确的估计结果。在压缩感知理论中恢复信号的算法主要有三类：凸松弛,贪婪算法和贝叶斯学习。贪婪算法的计算复杂度低,实现容易,在实时性要求较高的通信技术中更加适合作为信号恢复算法。本文将讨论贪婪算法作为信道估计的算法问题和优势,并从两个方面对贪婪算法进行改进。第一,讨论在信号稀疏度未知的情况下,如何使贪婪算法自适应停止迭代。本文中通过对OMP算法中的剩余向量进行分析,然后采用信号检测理论来给出算法停止迭代检测器。第二,最后本文将从贝叶斯的角度出发,改进贪婪算法的评估剩余向量与矩阵相关度的标准。通过计算复杂度分析和数值仿真来显示出以上两个改进的有效性。
[Abstract]:Compressed perception is one of the most important theories put forward in this century. It claims that sparse signals can be recovered from fewer sample points than Nyquist sampling rate. Compression sensing theory is a data acquisition and compression technology developed under the background of image processing. It not only plays an important role in image processing, but also in other fields, such as channel estimation, signal sampling and data compression. The theory of compressed sensing has great application value. This paper will mainly introduce the basic theory of compressed sensing and the application of compressed sensing in channel estimation. In the wireless communication environment with more obstacles, the transmitted signal usually reaches the receiver through several different delay paths. The multi-channel wireless channel is usually modeled as a linear filter. The received signal is a linear superposition of the transmitted signal with different attenuation and delay. The measured data show that the impulse response of the transmission path can be regarded as an approximate sparse signal in time domain. Compression sensing technique can be used to estimate the channel, because the sparsity of the channel is considered. Therefore, more accurate estimation results can be obtained than traditional channel estimation methods. There are three kinds of algorithms to recover signals in compression sensing theory: convex relaxation. Greedy algorithm and Bayesian learning. Greedy algorithm has low computational complexity and is easy to implement. In the real-time communication technology is more suitable as a signal recovery algorithm. This paper will discuss the greedy algorithm as channel estimation algorithm and advantages, and improve the greedy algorithm from two aspects. First. This paper discusses how to make greedy algorithm self-adaptively stop iteration when the signal sparsity is unknown. In this paper, the residual vectors in OMP algorithm are analyzed. Then the signal detection theory is used to give the algorithm to stop the iterative detector. Second, this paper will start from the perspective of Bayes. The computational complexity analysis and numerical simulation are used to show the effectiveness of these two improvements.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.23

【相似文献】