Please use this identifier to cite or link to this item:
標題: 具有「不相關之帶通干擾」之主動噪音控制系統的開發
Development of Active Noise Control Systems with Uncorrelated Pass-Band Disturbance
作者: 林忠逸
關鍵字: 應用研究;Active Noise Control;電信工程;主動噪音控制;不相關之干擾;適應性演算法;FIR 率波器;置換誤差;收斂性分析;盲訊號分離;Uncorrelated Disturbance;Adaptive Algorithms;FIRFilters;Commutation Error;Convergence Analysis;Blind Source Separation
本研究計畫擬在主動噪音控制系統中加入不相關之帶通干擾( uncorrelated pass-band disturbance) 的考量, 設計新的適應性消音演算法,降低或排除干擾的影響,並提升消音性能。本研究擬以三年的時間來進行,並考慮兩種量測方式,包括:( 1)有獨立的參考訊號可用,及( 2)無獨立的參考訊號可用。首先在有獨立參考訊號可用的系統中,擬多使用一個適應性的FIR 率波器來取得干擾之估測值,然後將其與置換誤差同時考量於適應性消音演算法的設計中。由於將存在著兩個適應性演算法,如何調整個別演算法的步進大小( step size)以確保合成之演算法的收斂性,必須先解決。第一年擬針對合成的演算法來建立收斂性的學理分析方法,以確保收斂性,並以電腦模擬來探討其在主動噪音控制應用上的性能。第二年擬利用第一年的成果來進行主動噪音控制的實驗與比較,並加入正規化步進大小( normalized step sizes)的研究與分析。第三年將考慮無獨立參考訊號可用的主動噪音控制系統,擬採用盲訊號分離( BSS)演算法來取代上述之適應性FIR 率波器,這是因為盲訊號分離演算法可以在只知道獨立訊號源數目的情況下從混合的訊號中將獨立的訊號源分離出來;擬用分離出來的訊號來設計與發展適應性消音演算法,降低干擾的影響,並以電腦模擬或實驗來驗證與討論。

In this study, uncorrelated pass-band disturbance is included in active noise control(ANC) systems to develop new adaptive algorithms so that effect of the disturbance can beattenuated and ANC performance can be improved. Two different types of ANC systems willbe considered in this three-year study including ANC systems with and without independentreference signal. In the first year, we shall consider ANC systems with an independentreference signal and propose to use an additional adaptive FIR filter to obtain an estimate ofthe disturbance. This estimate will be applied together with a commutation error to deriveanother adaptive algorithm for ANC. A combined ANC algorithm can thus be formed fromthese two adaptive algorithms with different step sizes. A deterministic method to analyzeconvergence property of the combined ANC algorithm will then be established such thatconvergence of the algorithm can be ensured. In addition, computer simulation of thealgorithm will be conducted to examine ANC performance in terms of convergence rate andlevel of noise reduction. In the second year, ANC experiments of the combined ANCalgorithm will be set up and carried out based on the simulation results of the first year.Moreover, normalized step sizes will be studied and compared. In the third year, we shallconsider ANC systems without any independent reference signal. Since blind sourceseparation (BSS) algorithm can separate independent source signals from a mixture signalwithout any prior information except for the number of independent sources, it will be usedto replace the adaptive FIR filter to obtain an estimate of the disturbance from outputs of twomicrophones. Results of the BSS algorithm will then be used to derive ANC algorithms suchthat effect of the disturbance can be reduced. Computer simulation and/or experiment will beconduced to investigate the effectiveness of the proposed BSS-based ANC approach.
其他識別: NSC99-2221-E005-061
Appears in Collections:機械工程學系所

Show full item record

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.