基于高阶累积量的声矢量阵列信号DOA估计算法研究

发布时间：2018-01-21 10:49

本文关键词： 声矢量传感器阵列高阶累积量高斯色噪声 DOA估计 MUSIC ESPRIT　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：声矢量传感器是近年兴起的一种新型声音信号探测设备。它能够同步共点的测量某地的声压信息和三维振速信息,相较于传统的声压传感器,不仅能够捕获更多的有用信息,而且探测更加灵敏,探测范围更加广泛。在通信、雷达、定位、生物医学及军事国防方面,正逐步取代传统的声压传感器。因此,无论从国民经济,还是国家安全领域来看,对声矢量传感器的研究都具有十分重要的意义。目前对声矢量传感器技术的研究主要两大部分:其一是声矢量传感器机械构成的研究;另一个是对声矢量传感器的应用方面的研究,即对声矢量传感器所接收的信号进行处理。如果说前者是骨架的构建,后者便是赋予这副骨架灵魂。本文将侧重于对后者的研究。波达方向估计(Direction Of Arrival,DOA),作为阵列信号处理的重要课题之一,也是声矢量传感器阵列实际应用中面临的重要难题。能够准确的估计出信号的到达角,是后续声矢量传感器的应用基础。自声矢量传感器问世以来,许多基于二阶统计量的算法,如声矢量MUSIC算法(Acoustic Vector Sensor MUSIC,A-MUSIC),声矢量ESPRIT(Acoustic Vector Sensor ESPRIT,A-ESPRIT)等,这些算法虽然对高斯白噪声有一定的抑制作用,但是对高斯色噪声抑制却无能为力。考虑实际工况中多是高斯色噪声,该问题的存在严重制约了声矢量技术向实际应用的转化。高阶累积量,由于其对高斯噪声的不敏感性,一经提出便受到了许多专家学者的青睐,各种基于常规阵元的高阶累积量算法层出不穷,并取得了很好的效果,很多已经付诸于实际应用。但是对于声矢量传感器而言,高阶累积量的处理却鲜有人为。究其原因,莫过于声矢量传感器的特性导致用高阶累积量的处理计算复杂、计算量大。基于上述原因,本文对基于高阶累积量的声矢量阵列DOA估计问题进行了研究。首先,从抑制高斯色噪声的角度出发,推导了声矢量阵列的高阶累积量DOA估计算法。其次,考虑到实际应用中计算量的问题,提出了基于高阶累积量对角切片的声矢量MUSIC算法。再次,为了取得更加精确的DOA估计效果,提出了基于累积量3,4维切片的声矢量MUSIC算法。接着,为了更进一步减少计算量,将ESPRIT算法引入到高阶累积量切片中,提出了声矢量阵列高阶累积量对角切片的ESPRIT算法。然后,为了克服高阶累积量对角切片ESPRIT计算中导致DOA估计精度降低的问题,本文提出了改进的声矢量阵列高阶累积量对角切片ESPRIT算法。最后,通过MATLAB对上述的算法进行了仿真验证,结果表明所提算法具有较好的DOA估计性能。本文主要针对声矢量传感器阵列的DOA估计进行了研究,以期望有助于声矢量阵列信号的理论研究和实际应用。
[Abstract]:Acoustic vector sensor is a new kind of sound signal detection equipment, which can synchronously measure the sound pressure information and the three-dimensional vibration velocity information of a place at the same time, compared with the traditional sound pressure sensor. Not only can more useful information be captured, but detection is more sensitive and more extensive. In communications, radar, positioning, biomedical and military defense. Is gradually replacing the traditional sound pressure sensor. Therefore, whether from the national economy or the field of national security. The research of acoustic vector sensor is of great significance. At present, there are two main parts in the research of acoustic vector sensor technology: one is the research on the mechanical structure of acoustic vector sensor; The other is the research on the application of acoustic vector sensor, that is, processing the signal received by the acoustic vector sensor, if the former is the construction of skeleton. The latter is to give the skeleton soul. This paper will focus on the study of the latter. As an important subject of array signal processing, it is also an important problem in the practical application of acoustic vector sensor array. It can accurately estimate the arrival angle of the signal. Since the advent of acoustic vector sensors, many algorithms based on second-order statistics have been developed. For example, acoustic Vector Sensor MUSICI A-MUSICA (acoustic vector MUSIC algorithm). Acoustic vector ESPRIT(Acoustic Vector Sensor Esprit (A-Esprit), et al. Although these algorithms can suppress Gao Si's white noise to some extent, there is no way to suppress Gao Si's color noise. The existence of this problem seriously restricts the transformation of acoustic vector technology to practical application. Because of its insensitivity to Gao Si noise, high order cumulant is favored by many experts and scholars. A variety of high-order cumulant algorithms based on conventional array elements have emerged and achieved good results, many of which have been applied to practical applications, but for acoustic vector sensors. The processing of high-order cumulants is rare. The reason is that the characteristics of acoustic vector sensors lead to the complexity and complexity of the processing of high-order cumulants. In this paper, the DOA estimation problem of acoustic vector array based on higher-order cumulant is studied. Firstly, from the point of view of suppressing Gao Si color noise. The high-order cumulant DOA estimation algorithm for acoustic vector array is derived. Secondly, considering the computational complexity in practical application, a novel acoustic vector MUSIC algorithm based on diagonal slice of high-order cumulant is proposed. In order to obtain more accurate DOA estimation effect, an acoustical vector MUSIC algorithm based on cumulant 3-dimensional slice is proposed. Then, in order to further reduce the computational complexity. The ESPRIT algorithm is introduced into higher-order cumulant slice, and the ESPRIT algorithm for diagonal slice of high-order cumulant in acoustic vector array is proposed. In order to overcome the problem of reducing the accuracy of DOA estimation in diagonal slice ESPRIT calculation of high-order cumulants. In this paper, an improved ESPRIT algorithm of high order cumulant diagonal slicing of acoustic vector array is proposed. Finally, the above algorithm is simulated by MATLAB. The results show that the proposed algorithm has good DOA estimation performance. This paper mainly focuses on the DOA estimation of acoustic vector sensor array. It is expected to be helpful to the theoretical research and practical application of acoustic vector array signals.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212

【相似文献】