本文關(guān)鍵詞： 壓縮感知 信道估計(jì) 稀疏性 信號(hào)檢測(cè) 貪婪算法　出處：《電子科技大學(xué)》2014年碩士論文　論文類(lèi)型：學(xué)位論文

【摘要】：壓縮感知是在本世紀(jì)所提出最重要的理論之一,它聲稱(chēng)可以將稀疏信號(hào)從比奈奎斯特采樣率得到地更少的樣本點(diǎn)中恢復(fù)出來(lái)。壓縮感知理論是在圖像處理的背景下,發(fā)展出來(lái)的數(shù)據(jù)采集和壓縮技術(shù),革新了傳統(tǒng)對(duì)信號(hào)的采集和存儲(chǔ)的手段。因此它不僅在圖像處理方面有著重要作用,在其他一些領(lǐng)域,如信道估計(jì),信號(hào)采樣以及數(shù)據(jù)壓縮等,壓縮感知理論都有極大的應(yīng)用價(jià)值。本文將主要介紹壓縮感知的基本理論以及壓縮感知在信道估計(jì)中的應(yīng)用。在無(wú)線(xiàn)通信中,在障礙物物較多的無(wú)線(xiàn)通信環(huán)境下,發(fā)送信號(hào)通常經(jīng)過(guò)多條不同延時(shí)的路徑到達(dá)接收者。多經(jīng)無(wú)線(xiàn)信道通常被建模為一個(gè)線(xiàn)性濾波器,接收到的信號(hào)為發(fā)送信號(hào)經(jīng)不同衰減和延遲的線(xiàn)性疊加。實(shí)測(cè)數(shù)據(jù)表明,傳輸路徑的沖激響應(yīng)在時(shí)域上可看作是一個(gè)近似稀疏的信號(hào)。因此,可以采用壓縮感知技術(shù)對(duì)信道進(jìn)行估計(jì)。由于考慮了信道的稀疏性,所以可以得到比傳統(tǒng)信道估計(jì)方法更準(zhǔn)確的估計(jì)結(jié)果。在壓縮感知理論中恢復(fù)信號(hào)的算法主要有三類(lèi)：凸松弛,貪婪算法和貝葉斯學(xué)習(xí)。貪婪算法的計(jì)算復(fù)雜度低,實(shí)現(xiàn)容易,在實(shí)時(shí)性要求較高的通信技術(shù)中更加適合作為信號(hào)恢復(fù)算法。本文將討論貪婪算法作為信道估計(jì)的算法問(wèn)題和優(yōu)勢(shì),并從兩個(gè)方面對(duì)貪婪算法進(jìn)行改進(jìn)。第一,討論在信號(hào)稀疏度未知的情況下,如何使貪婪算法自適應(yīng)停止迭代。本文中通過(guò)對(duì)OMP算法中的剩余向量進(jìn)行分析,然后采用信號(hào)檢測(cè)理論來(lái)給出算法停止迭代檢測(cè)器。第二,最后本文將從貝葉斯的角度出發(fā),改進(jìn)貪婪算法的評(píng)估剩余向量與矩陣相關(guān)度的標(biāo)準(zhǔn)。通過(guò)計(jì)算復(fù)雜度分析和數(shù)值仿真來(lái)顯示出以上兩個(gè)改進(jìn)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.
【學(xué)位授予單位】：電子科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類(lèi)號(hào)】：TN911.23

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 楊潔;;基于貪婪算法的衛(wèi)星區(qū)域觀(guān)測(cè)擺角方案選擇方法[J];廣西科學(xué)院學(xué)報(bào);2006年02期

2 申時(shí)凱;吳紹兵;申浩如;王付艷;管彥慶;;計(jì)算最短公共超串的貪婪算法[J];計(jì)算機(jī)工程與設(shè)計(jì);2007年08期

3 王杰;剛軼金;李鳳光;吳偉巍;;改進(jìn)貪婪算法在博客突發(fā)事件檢測(cè)中的研究[J];計(jì)算機(jī)工程與應(yīng)用;2008年34期

4 周柳陽(yáng);高珩;梁翥;;貪婪算法的實(shí)際應(yīng)用[J];硅谷;2009年02期

5 李e，

本文編號(hào)：1446324