基于子空间分析的DOA估计算法研究

发布时间：2018-03-08 16:17

本文选题：阵列信号处理　切入点：二维DOA估计　出处：《南京邮电大学》2017年硕士论文　论文类型：学位论文

【摘要】：空间信号波达方向(DOA)估计是阵列信号处理的一个重要分支。基于子空间分析的二维DOA估计能够实现对空间信号更加准确的定位,并且充分利用阵列数据矩阵或其二阶统计量的内在结构特性,具有分辨率高、复杂度低、实时性好等优点,是DOA估计领域的研究重点之一。随着二维DOA估计在工程实践中的实用性研究不断深入,人们对算法的估计精度、分辨率、计算复杂度等要求越来越高。本文在现有子空间类估计算法研究成果的基础上,针对二维DOA估计提出了三种改进算法,主要工作如下:(1)为了提高面阵下二维DOA估计算法的性能,本文提出了一种基于传播算子(PM)的改进算法。从面阵接收信号中按一定规则提取相互重合的四个子面阵,由子面阵接收数据间的互相关矩阵创建一个新的数据矩阵,再利用PM算法得到旋转不变关系矩阵,最后得到自动匹配的二维角度估计。本文算法的计算复杂度要小于传统面阵ESPRIT算法。仿真结果表明,在信噪比较低或快拍数较小时,由于复用了互相关矩阵中的数据,改进算法的角度估计精度要好于面阵ESPRIT算法。(2)对阵列孔径进行扩展是提高DOA估计性能的一个重要手段。本文在双平行线阵模型下提出了一种高效的二维测向改进算法。主要利用阵列流形矩阵的共轭对称特性来扩展阵列的有效孔径,并结合PM算法来实现自动配对的二维参数估计。仿真结果表明,与现有的一些测向算法相比,本文算法的角度估计精度显著提高,且复杂度较低。(3)在相干信源DOA估计问题上,本文基于L型阵列提出了一种改进的相干信源二维DOA估计算法。改进算法提取互相关矩阵的第一列,利用相关规则构建一个列秩与信源相关性无关的新矩阵,然后对其正交化,构造求根多项式来降低计算复杂度。仿真结果表明,与现有的一些解相干算法相比,本文算法扩展了阵列有效孔径,提高了角度估计精度。
[Abstract]:Spatial signal DOA estimation is an important branch of array signal processing. Two-dimensional DOA estimation based on subspace analysis can locate the spatial signal more accurately. And make full use of the inherent structure characteristic of array data matrix or its second order statistics, it has the advantages of high resolution, low complexity, good real time and so on. It is one of the key points in the field of DOA estimation. With the further research on the practicability of two-dimensional DOA estimation in engineering practice, the estimation accuracy and resolution of the algorithm have been studied. Based on the existing research results of subspace class estimation algorithms, this paper proposes three improved algorithms for two-dimensional DOA estimation, the main work of which is as follows: (1) in order to improve the performance of two-dimensional DOA estimation algorithm under plane array, In this paper, an improved algorithm based on propagation operator PM) is proposed. Four overlapping subarrays are extracted from the received signals of the array according to certain rules, and a new data matrix is created by the cross-correlation matrix between the received data of the subarray. Then the rotation invariant matrix is obtained by using PM algorithm, and the 2-D angle estimation of automatic matching is obtained. The computational complexity of this algorithm is less than that of the traditional ESPRIT algorithm. The simulation results show that, when the SNR is low or the number of beats is small, Because the data in the cross-correlation matrix is multiplexed, The angle estimation accuracy of the improved algorithm is better than that of the plane array ESPRIT algorithm. The expansion of the array aperture is an important means to improve the performance of the DOA estimation. In this paper, an efficient two-dimensional direction finding improvement based on the dual parallel linear array model is proposed. The conjugate symmetry of the array manifold matrix is used to expand the effective aperture of the array. The simulation results show that compared with some existing directional finding algorithms, the angle estimation accuracy of this algorithm is significantly improved, and the complexity is lower. 3) in the problem of coherent source DOA estimation, the simulation results show that the proposed algorithm can be used to estimate the DOA of coherent source. In this paper, an improved two-dimensional DOA estimation algorithm for coherent sources is proposed based on L-type arrays. The improved algorithm extracts the first column of the cross-correlation matrix, constructs a new matrix with column rank independent of source correlation using correlation rules, and then orthogonalizes it. The simulation results show that the proposed algorithm extends the effective aperture of the array and improves the precision of angle estimation.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN911.7

【参考文献】