基于协方差矩阵高阶幂的二维DOA估计新算法

A Novel 2DDOA Estimation Algorithm Based on High-order Power of Covariance Matrix

针对稀疏分解方法进行均匀圆阵(UCA)的二维波达方向(DOA)估计运算复杂度大的问题,提出了一种基于协方差矩阵高阶幂稀疏分解的二维DOA估计新算法。该算法首先利用协方差矩阵高阶幂无需进行特征值分解和信源数估计的特性,构建了协方差矩阵高阶幂的稀疏分解向量;然后运用粒度分层思想,构造了粗区域估计和细方位估计的分层多粒度的快速分解模型,分层字典的长度大大减少,在保持估计精度的前提下,算法运算时间远小于现有的恒定冗余字典的稀疏分解方法,从而解决了基于稀疏分解的圆阵二维DOA估计问题。论文提出的算法与二维MUSIC算法相比,估计精度高,且能满足对相干信号的估计。仿真结果验证了算法的有效性和可行性。

To reduce the computational complexity in the sparse decomposition of 2D direction of arrival （DOA） estimation based on uniform circular arrays （UCAs）, a novel 2D DOA estimation algorithm using sparse decomposition of higherorder power of covariance matrix was proposed. First, this method avoids the estimation of the number of signals and eigenvalue decompsition through using the highorder power of a covariance matrix as the vector of sparse decomposition. Then a fast region direction detection decomposition algorithm based on the hierarchical granularity model is put forward. This new method can construct a redundant dictionary adaptively based on the distribution of space signals, thus reducing the computational load greatly while still maintaining high estimation accuracy. Compared with the 2D MUSIC method,this algorithm not only provides better 2D DOA performance but also possesses the capability of estimating coherent signals. Simulation results confirm its effectiveness and feasibility.