一種基于聲陣列信息融合及改進EEMD的信號降噪方法
發(fā)布時間:2018-09-19 12:19
【摘要】:針對聲陣列多通道信號的去噪問題,提出一種基于多傳聲器信息融合輔助的改進總體平均經(jīng)驗模態(tài)分解(Ensemble Empirical Mode Decomposition,EEMD)的被動聲信號去噪方法。對標準EEMD進行改進,通過多通道信號頻譜分析,選取多傳聲器信號最小有效頻率作為各通道信號EEMD分解的篩選截止頻率,采用改進的EEMD算法將原始信號快速分解為完備的IMF分量,有效抑制了模態(tài)混疊現(xiàn)象并提高信號分解效率;引入聲陣列時延矢量封閉準則(Time Delay Vector Close Rule,TDVCR)概念,結合多傳聲器數(shù)據(jù)一致性融合及信號相關性理論,對各IMF分量進行相應的權重計算,再由已確定權值對各IMF分量進行加權重構得到去噪信號;最終通過半實物仿真試驗以及同傳統(tǒng)EMD去噪的比較驗證了該算法在多通道信號去噪中的有效性和實用性。
[Abstract]:A passive acoustic signal de-noising method based on multi-microphone information fusion assisted by improved population average empirical mode decomposition (Ensemble Empirical Mode Decomposition,EEMD) is proposed for the de-noising of acoustic array multi-channel signals. The standard EEMD is improved. The minimum effective frequency of the multi-microphone signal is selected as the screening cutoff frequency of the EEMD decomposition of each channel signal through the multi-channel signal spectrum analysis. The improved EEMD algorithm is used to decompose the original signal into complete IMF components quickly, which effectively reduces the mode aliasing phenomenon and improves the signal decomposition efficiency. The concept of acoustic array time-delay vector closure criterion (Time Delay Vector Close Rule,TDVCR) is introduced. Combining the data consistency fusion of multi-microphone and the theory of signal correlation, the corresponding weight of each IMF component is calculated, and then the de-noising signal is obtained by the weighted reconstruction of each IMF component with the determined weight value. Finally, the effectiveness and practicability of the algorithm in multi-channel signal denoising are verified by the hardware-in-the-loop simulation experiment and the comparison with the traditional EMD de-noising algorithm.
【作者單位】: 南京理工大學機械工程學院;貴州大學智能信息處理研究所;
【基金】:國家自然科學基金(61263005)
【分類號】:TB535
本文編號:2250105
[Abstract]:A passive acoustic signal de-noising method based on multi-microphone information fusion assisted by improved population average empirical mode decomposition (Ensemble Empirical Mode Decomposition,EEMD) is proposed for the de-noising of acoustic array multi-channel signals. The standard EEMD is improved. The minimum effective frequency of the multi-microphone signal is selected as the screening cutoff frequency of the EEMD decomposition of each channel signal through the multi-channel signal spectrum analysis. The improved EEMD algorithm is used to decompose the original signal into complete IMF components quickly, which effectively reduces the mode aliasing phenomenon and improves the signal decomposition efficiency. The concept of acoustic array time-delay vector closure criterion (Time Delay Vector Close Rule,TDVCR) is introduced. Combining the data consistency fusion of multi-microphone and the theory of signal correlation, the corresponding weight of each IMF component is calculated, and then the de-noising signal is obtained by the weighted reconstruction of each IMF component with the determined weight value. Finally, the effectiveness and practicability of the algorithm in multi-channel signal denoising are verified by the hardware-in-the-loop simulation experiment and the comparison with the traditional EMD de-noising algorithm.
【作者單位】: 南京理工大學機械工程學院;貴州大學智能信息處理研究所;
【基金】:國家自然科學基金(61263005)
【分類號】:TB535
【相似文獻】
相關期刊論文 前1條
1 林向真,明代都;SF-10深海復合棒振子的設計及其布陣應用[J];聲學學報;1979年03期
相關會議論文 前1條
1 陳寶柱;;水聲陣元域標準測試數(shù)據(jù)庫構建研究[A];中國聲學學會水聲學分會2013年全國水聲學學術會議論文集[C];2013年
相關博士學位論文 前1條
1 蔣佳佳;水聲陣列探測算法及系統(tǒng)關鍵技術研究[D];天津大學;2014年
相關碩士學位論文 前1條
1 李會超;聲陣列超指向性測量機理及方法研究[D];青島理工大學;2016年
,本文編號:2250105
本文鏈接:http://sikaile.net/guanlilunwen/gongchengguanli/2250105.html
最近更新
教材專著