EMD去噪与MUSIC算法在DOA估计中的联合应用

发布时间：2019-01-07 20:46

【摘要】：矢量水听器与传统的声压水听器相比具有很多的优势,而由其组成的矢量阵更能提高阵列的性能。MUSIC算法是阵列信号处理中具有里程碑意义的一类算法,但它的缺陷在于性能严重依赖于信噪比,其分辨性能会随着信噪比的降低而显著下降。本文提出了一种基于EMD去噪与MUSIC算法相结合的联合估计方法。其基本思路是：首先用本文构造的基于噪声统计特性的EMD去噪方法对各基元接收到的信号分别进行EMD去噪预处理,再将去噪后的数据进行MUSIC方位估计。本文前半部分研究了EMD算法以及其在信号去噪中的应用。文中深入探讨了噪声及噪声经EMD分解后各个分量的统计特性。在假定首个模态分量为噪声的前提下,提出了一种基于噪声统计特性的EMD去噪新方法。首先,将经过EMD分解得到的首个模态分量每次随机排序并与其他的分量进行重构,将多次重构结果累加求平均后,由于噪声的随机特性,便得到了信噪比改善的信号分量：而后再对新分量重新进行分解,重复该操作数次；最后,得到噪声功率得到很大抑制的信号分量。该方法可以简化为：随机排序-重构-累加-求平均-再分解-重复前述操作的过程。仿真实验表明,该方法在低信噪比的信号去噪中表现了良好的性能,其为低信噪比去噪提供了新思路。本文后半部分深入研究了MUSIC算法的基本理论,并分析了影响MUSIC算法定向精度的各个因素。针对MUSIC算法在低信噪比时无法准确对信号源进行定向及对多目标的分辨能力差的问题,结合前半部改进后的基于噪声统计特性的EMD去噪新方法,通过仿真实验表明,MUSIC空间谱的主波束宽度锐化,旁瓣降低,可提高对信号源的定向精度及对多目标的分辨能力。
[Abstract]:The vector hydrophone has many advantages compared with the traditional acoustic hydrophone, and the vector array composed of vector hydrophone can improve the performance of the array. MUSIC algorithm is a kind of landmark algorithm in array signal processing. But its performance depends heavily on signal-to-noise ratio (SNR), and its resolution performance will decrease with the decrease of SNR. In this paper, a joint estimation method based on EMD denoising and MUSIC algorithm is proposed. The basic ideas are as follows: firstly, the EMD denoising method based on noise statistics is used to pre-process the received signals by EMD, and then to estimate the MUSIC azimuth of the de-noised data. In the first half of this paper, the EMD algorithm and its application in signal denoising are studied. The statistical properties of each component of noise and noise decomposed by EMD are discussed in detail. Under the assumption that the first modal component is noise, a new EMD denoising method based on the statistical characteristics of noise is proposed. First of all, the first modal component obtained by EMD decomposition is sorted randomly each time and reconstructed with other components. After the multiple reconstruction results are accumulated to average, because of the random characteristic of noise, The signal component with improved signal-to-noise ratio (SNR) is obtained: then the new component is decomposed again and the operation is repeated several times; Finally, the noise power is greatly suppressed. The method can be simplified as: random sort-refactoring-accumulative-average-refactoring-repeating the previous operation. The simulation results show that the proposed method performs well in signal denoising with low signal-to-noise ratio (SNR) and provides a new idea for de-noising with low signal-to-noise ratio (SNR). In the second half of this paper, the basic theory of MUSIC algorithm is deeply studied, and the factors that affect the orientation accuracy of MUSIC algorithm are analyzed. Aiming at the problem that the MUSIC algorithm can not accurately orient the signal source at low signal-to-noise ratio (SNR) and has poor resolution to multiple targets, combined with the improved EMD denoising method based on the noise statistical characteristics in the first half, the simulation results show that the proposed method can be used to solve the problem. The main beam width of MUSIC spectrum is sharpened and the sidelobe is reduced, which can improve the orientation accuracy of signal source and the resolution of multi-target.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN911.4

【参考文献】