特殊阵列下的快速波达方向估计
发布时间：2018-11-07 17:14
【摘要】：作为阵列信号处理的重要分支之一,波达方向(Direction of Arrival,DOA)估计技术具有广泛的应用背景,其相关研究也一直是阵列信号处理领域的热点。从问世至今,DOA估计经历了从高分辨到超分辨的发展历程,在估计性能日益完善的同时,如何降低计算量以便超分辨估计算法由理论走向工程应用逐渐成为另一个热点问题。本文以均匀线阵(Uniform Linear Array,ULA)和互质阵为依托,针对常规波达方向估计计算量大的问题,提出三种降低计算量的快速算法,为超分辨估计算法工程化提供理论参考。本文的主要研究内容如下:首先,在充分研究基于L阵二维DOA估计的CESA算法后,提出了一种新的基于ULA的一维快速DOA估计算法。新算法在阵型上将ULA划分成两个阵元数相等的子阵,然后求子阵的前、后向互协方差矩阵并构造联合互协方差矩阵,再对该矩阵的第一列向量线性操作以获取信号子空间,最后构造多项式求根求得DOA。仿真实验证明,该算法在保证估计精度可接受的同时有效降低了协方差矩阵计算和子空间分解的计算量。其次,在深入研究针对L阵二维DOA估计的CODE算法后,提出一种新的基于ULA的子空间快速估计算法。新算法先将ULA划分成两个子阵,然后将流型矩阵划分成两个存在旋转不变关系的子矩阵,根据子阵互协方差矩阵与流型矩阵的对应关系,将互协方差矩阵也划分成满足旋转不变性的两个子矩阵,最后利用这两个子矩阵求解出旋转不变关系矩阵,继而求出子空间、构造多项式求解DOA。新算法相对于CODE算法实现了估计精度的提高,同时有效降低了子空间获取的计算量。最后,针对现阶段较为热点的互质阵提出了一种新的DOA快速估计算法,新算法利用互质阵部分均匀线性的特点将互质阵看成由两个ULA子阵构成,然后分别根据root-MUSIC算法构造子阵多项式,之后利用子阵多项式构造互质阵的多项式,最后研究证明该互质阵多项式单位圆上的根就是待求的DOA对应的根。仿真实验证明,新算法的估计精度和效率均高于求根MUSIC算法及现有的同类算法。
[Abstract]:As an important branch of array signal processing, DOA (Direction of Arrival,DOA) estimation technology has a wide application background, and its related research has been a hot spot in the field of array signal processing. Up to now, DOA estimation has gone through the development from high resolution to super resolution. While the performance of DOA estimation is becoming more and more perfect, how to reduce the computational complexity so that the super-resolution estimation algorithm can be applied from theory to engineering has gradually become another hot issue. In this paper, based on the uniform linear array (Uniform Linear Array,ULA) and the mutual mass matrix, three fast algorithms to reduce the computational complexity are proposed to solve the problem of large computational complexity in the conventional DOA estimation, which provides a theoretical reference for the engineering of the super-resolution estimation algorithm. The main contents of this paper are as follows: firstly, after fully studying the CESA algorithm based on L-matrix two-dimensional DOA estimation, a new one-dimensional fast DOA estimation algorithm based on ULA is proposed. The new algorithm divides ULA into two submatrices with equal number of elements in the matrix, then obtains the forward and backward cross-covariance matrix of the submatrix and constructs the joint cross-covariance matrix. Then the first column vector of the matrix is linearly operated to obtain the signal subspace. Finally, we construct polynomial to find the root of DOA.. Simulation results show that the proposed algorithm can effectively reduce the computational complexity of covariance matrix calculation and subspace decomposition while ensuring acceptable estimation accuracy. Secondly, after deeply studying the CODE algorithm for 2-D DOA estimation of L-matrix, a new fast subspace estimation algorithm based on ULA is proposed. The new algorithm first divides ULA into two submatrices, then divides the flow pattern matrix into two submatrices with rotation-invariant relations. According to the corresponding relationship between the submatrix cross-covariance matrix and the flow pattern matrix, the new algorithm divides the flow pattern matrix into two submatrices. The cross covariance matrix is also divided into two submatrices which satisfy the rotation invariance. Finally, the rotation invariant relation matrix is solved by using these two submatrices, and then the subspace is obtained, and the polynomial is constructed to solve the DOA.. Compared with the CODE algorithm, the new algorithm improves the estimation accuracy and reduces the computational complexity of subspace acquisition. Finally, a new DOA fast estimation algorithm is proposed for the hot mutual-prime matrix. The new algorithm takes advantage of the partial uniform linearity of the mutual-prime matrix to treat the matrix as two ULA subarrays. Then the submatrix polynomials are constructed according to the root-MUSIC algorithm, and then the polynomials of the coprime matrix are constructed by using the submatrix polynomials. Finally, it is proved that the roots on the unit circle of the coprime polynomials are the DOA corresponding roots to be solved. The simulation results show that the estimation accuracy and efficiency of the new algorithm are higher than that of the root seeking MUSIC algorithm and the existing similar algorithms.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN911.7
本文编号：2317024
[Abstract]:As an important branch of array signal processing, DOA (Direction of Arrival,DOA) estimation technology has a wide application background, and its related research has been a hot spot in the field of array signal processing. Up to now, DOA estimation has gone through the development from high resolution to super resolution. While the performance of DOA estimation is becoming more and more perfect, how to reduce the computational complexity so that the super-resolution estimation algorithm can be applied from theory to engineering has gradually become another hot issue. In this paper, based on the uniform linear array (Uniform Linear Array,ULA) and the mutual mass matrix, three fast algorithms to reduce the computational complexity are proposed to solve the problem of large computational complexity in the conventional DOA estimation, which provides a theoretical reference for the engineering of the super-resolution estimation algorithm. The main contents of this paper are as follows: firstly, after fully studying the CESA algorithm based on L-matrix two-dimensional DOA estimation, a new one-dimensional fast DOA estimation algorithm based on ULA is proposed. The new algorithm divides ULA into two submatrices with equal number of elements in the matrix, then obtains the forward and backward cross-covariance matrix of the submatrix and constructs the joint cross-covariance matrix. Then the first column vector of the matrix is linearly operated to obtain the signal subspace. Finally, we construct polynomial to find the root of DOA.. Simulation results show that the proposed algorithm can effectively reduce the computational complexity of covariance matrix calculation and subspace decomposition while ensuring acceptable estimation accuracy. Secondly, after deeply studying the CODE algorithm for 2-D DOA estimation of L-matrix, a new fast subspace estimation algorithm based on ULA is proposed. The new algorithm first divides ULA into two submatrices, then divides the flow pattern matrix into two submatrices with rotation-invariant relations. According to the corresponding relationship between the submatrix cross-covariance matrix and the flow pattern matrix, the new algorithm divides the flow pattern matrix into two submatrices. The cross covariance matrix is also divided into two submatrices which satisfy the rotation invariance. Finally, the rotation invariant relation matrix is solved by using these two submatrices, and then the subspace is obtained, and the polynomial is constructed to solve the DOA.. Compared with the CODE algorithm, the new algorithm improves the estimation accuracy and reduces the computational complexity of subspace acquisition. Finally, a new DOA fast estimation algorithm is proposed for the hot mutual-prime matrix. The new algorithm takes advantage of the partial uniform linearity of the mutual-prime matrix to treat the matrix as two ULA subarrays. Then the submatrix polynomials are constructed according to the root-MUSIC algorithm, and then the polynomials of the coprime matrix are constructed by using the submatrix polynomials. Finally, it is proved that the roots on the unit circle of the coprime polynomials are the DOA corresponding roots to be solved. The simulation results show that the estimation accuracy and efficiency of the new algorithm are higher than that of the root seeking MUSIC algorithm and the existing similar algorithms.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN911.7
【引证文献】
相关期刊论文 前1条
1 闫锋刚;荣加加;刘帅;沈毅;金铭;;联合互协方差矩阵的快速波达方向估计[J];系统工程与电子技术;2018年04期
，本文编号：2317024
本文链接：https://www.wllwen.com/kejilunwen/xinxigongchenglunwen/2317024.html
