基于一阶统计量的子空间旋转不变解相干算法

ESPRIT Decorrelation Algorithm Based on First Order Statistics

针对低信噪比和少快拍数情况下相干信源二维波达方向(direction of arrival,DOA)估计精度大幅下降的问题,提出了基于一阶统计量的子空间旋转不变(ESPRIT)解相干算法。该算法首先对阵元接收数据进行求均值运算,利用每个阵元接收数据的一阶统计量构造三个Toeplitz形式伪协方差矩阵,实现相干信源的解相干;然后构造一个扩展协方差矩阵并对其进行一次奇异值分解即可实现相干信源的二维DOA估计。仿真结果表明:该算法在低信噪比和少快拍数下的估计性能优于空域平滑波达方向矩阵法,估计精度高,且避免了传统算法对信号协方差矩阵的处理过程,计算复杂度低,实时性高。

In order to solve the problem that the accuracy of two-dimensional（2-D）direction of arrival（DOA）estimation of coherent signal sources would decline greatly with low SNR and shot snapshots,a novel 2-D ESPRIT algorithm based on first order statistics was proposed.Firstly,average operation on the data gathered by arrays was taken and three Toeplitz pseudo covariance matrices were constructed for decorrelation by using the first order statistics of array observed data.Then,the 2-D DOA estimation of coherent signals could be solved by using singular values decomposition for extended covariance matrix.Simulation results demonstrated that,compared with spatial smoothing DOA matrix method,the proposed algorithm had higher resolution characteristics with low SNR and shot snapshots.Meanwhile,the algorithm had excellent real-time performance and low computational complexity by avoiding processing of covariance matrix of conventional algorithms.