摘要: |
针对低信噪比情况下的时延估计,将二次相关(SC)时延估计与经验模态分解(EMD)算法结合,提出了EMD重构二次相关时延估计方法。该方法针对EMD重构时本征模态函数的选择,将倒谱法和谱减法相结合,提出新的本征模态函数中有用信号主导分量和噪声主导分量的区分方案。研究结果表明:EMD重构二次相关法较传统二次相关法抗噪性能更优,更能锐化二次相关峰值;在非高斯有色噪声和高斯白噪声情况下,分别将准确估计时延的信噪比降低了4 dB和2 dB。 |
关键词: 时延估计 经验模态分解 二次相关 声音定位 |
DOI: |
|
基金项目:广西科学研究与技术开发计划项目(桂科攻1598008-29);基于OFDM的无线-光网络用户端上变频技术研究(ZD2014146) |
|
Second correlation time delay estimation based on empirical mode decomposition reconstruction |
ZHOU Zhuowei,SUN Shanlin,LI Yun,CHEN Pangsheng,SHEN Jingwen |
() |
Abstract: |
For the purpose of enhancing anti-noise capability of time delay estimation,a novel method named EMD-SC is proposed by applying Empirical Mode Decomposition(EMD)technique to Second Correlation(SC)method in this paper. In addition,the cepstrum method and the subtraction of frequency spectrum method are used to select the signal-led Intrinsic Mode functions,which are produced by EMD,for signal reconstruction.Comparison of the simulation results between EMD-SC method and traditional Second Correlation method at various signal-to-noise ratio(SNR) levels shows that the proposed method obtains a much sharper second correlation peak,and decreases the SNR of precise estimation by 4 dB and 2 dB in the colored non-Gaussian noise and Gaussian white noise. |
Key words: time delay estimation empirical mode decomposition second correlation sound location |