發(fā)布時間：2018-03-14 17:54

  本文選題：LMS算法　切入點：回聲消除　出處：《西安電子科技大學(xué)》2014年碩士論文　論文類型：學(xué)位論文

【摘要】：近年來,語音信號處理技術(shù)和通信技術(shù)發(fā)展得越來越快,人們對通信系統(tǒng)的通話質(zhì)量要求也變得越來越高。免提語音通話系統(tǒng)得到了人們的青睞,例如電話視頻會議、車載免提電話等。其中,聲學(xué)回聲是影響這類免提語音通話質(zhì)量的首要因素。因此,解決免提語音通話系統(tǒng)中的聲學(xué)回聲問題是人們研究和關(guān)注的熱點問題。很多人對自適應(yīng)濾波算法做了深入研究,并通過MATLAB進行了仿真,取得了一定的成果。但是,由于實際通信環(huán)境的復(fù)雜性,致使仿真實驗并不能完全模擬出真實的通話環(huán)境,仿真結(jié)果和實際結(jié)果往往存在著比較大的差別。本文在研究自適應(yīng)濾波器及其相關(guān)算法的基礎(chǔ)上,致力于構(gòu)建一個基于DSP的聲學(xué)回聲消除硬件平臺,并且模擬出實際通信系統(tǒng)中的真實通話環(huán)境,采集實際的語音信號和回聲信號進行回聲消除處理,為自適應(yīng)濾波技術(shù)的硬件實現(xiàn)與應(yīng)用做出進一步嘗試。本文采取理論分析、軟件仿真和硬件設(shè)計實現(xiàn)相結(jié)合的研究方法。在研究自適應(yīng)算法的基礎(chǔ)上,用軟件仿真驗證相關(guān)參數(shù)對自適應(yīng)算法性能的影響,并通過仿真將自適應(yīng)算法運用于自適應(yīng)濾波實驗中,通過回聲消除結(jié)果來對比算法的性能。最后基于DSP構(gòu)建聲學(xué)回聲消除的硬件平臺,完成自適應(yīng)濾波器的硬件實現(xiàn)。首先,本文通過對聲學(xué)回聲產(chǎn)生原因的分析,研究了回聲消除算法的基本原理,重點研究了LMS算法的推導(dǎo)過程和基本原理,并對影響LMS算法性能和指標的因素做了分析。然后,研究了幾種變步長算法,并對變步長算法和標準LMS算法做了比較。其次,本文分別對標準LMS算法和變步長LMS算法進行了仿真,并且比較了兩者之間存在的差異。最后,本文以德州儀器公司2000系列芯片TMS320F28335為核心,構(gòu)建了一個模擬真實通信系統(tǒng)的聲學(xué)回聲消除平臺,用實際通話語音和回聲做了測試,并且把實際測試結(jié)果和仿真結(jié)果做了對比。同時,由于實際通信系統(tǒng)中存在著單端通話和雙端通話的問題,所以本文探討了實際應(yīng)用中的雙端通話檢測技術(shù),并且實現(xiàn)了對該技術(shù)的應(yīng)用。另外,本文探討了立體聲環(huán)境中的回聲消除問題,分析了當(dāng)前的研究現(xiàn)狀和未來的發(fā)展趨勢。
[Abstract]:In recent years, the technology of voice signal processing and communication has developed more and more rapidly, and the quality of communication system has become more and more demanding. Among them, acoustic echo is the primary factor that affects the quality of this kind of hands-free voice communication. To solve the acoustic echo problem in hands-free voice communication system is a hot issue that people study and pay close attention to. Many people have done in-depth research on adaptive filtering algorithm and simulated it through MATLAB. However, some achievements have been made. Because of the complexity of the actual communication environment, the simulation experiment can not completely simulate the real communication environment. The simulation results and the actual results are often quite different. Based on the study of adaptive filter and its related algorithms, this paper is devoted to constructing a hardware platform of acoustic echo cancellation based on DSP. And simulate the real communication environment in the actual communication system, collect the actual voice signal and echo signal for echo cancellation processing, and make a further attempt for the hardware implementation and application of adaptive filtering technology. Based on the research of adaptive algorithm, the influence of relevant parameters on the performance of adaptive algorithm is verified by software simulation. Finally, the hardware platform of acoustic echo cancellation is constructed based on DSP, and the hardware implementation of adaptive filter is completed. Based on the analysis of the cause of acoustic echo generation, the basic principle of echo cancellation algorithm is studied, the derivation process and basic principle of LMS algorithm are emphatically studied, and the factors influencing the performance and index of LMS algorithm are analyzed. Several variable step size algorithms are studied and compared with the standard LMS algorithm. Secondly, the standard LMS algorithm and the variable step size LMS algorithm are simulated, and the differences between them are compared. Based on Texas Instruments 2000 series chip TMS320F28335, an acoustic echo cancellation platform for simulating real communication system is constructed, and tested with actual voice and echo. At the same time, because of the problem of single and double end call in the actual communication system, this paper discusses the dual end call detection technology in practical application. In addition, the problem of echo cancellation in stereo environment is discussed, and the present research situation and future development trend are analyzed.
【學(xué)位授予單位】：西安電子科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TN912.3

【參考文獻】

相關(guān)期刊論文 前1條

1 朱燦彬;毛玉泉;李曉楠;李冰;;VoIP回聲消除算法研究[J];艦船電子工程;2007年04期

，

本文編號：1612308