基于数据矩阵的多个窄带信号频率和波达方向同时估计

Data Matrix Based Simultaneous Estimation of Frequencies and DOAs of Multiple Narrow Band Signals

以矩阵奇异值分解的性能研究为基础 ,针对等距线阵提出了一种多个窄带信号波达方向和频率的同时估计方法。该方法用采集到的数据矩阵在最小二乘意义下低秩近似无噪数据矩阵 ,并根据信号的窄带假设构造了可用 ESPRIT型方法求解待测参数的矩阵结构 ,频率和方向可自动配对 ,计算量较小。计算机模拟实验证实了该方法所作近似的有效性 ,显示了其高分辨率特性和在短数据条件下的鲁棒性。

A novel algorithm that utilizes some nice properties of single value decomposition (SVD) and can simultaneously estimate the azimuths and frequencies of multiple narrow band signals impinging on a ULA is presented. By minimizing the Frobenius norm of the error matrix, noiseless data matrix is approximated by a matrix of the same rank. Some matrix structures processed by ESPRIT like method to estimate the unknown parameters are constructed based on narrow band assumption of the signals. The azimuths and the frequencies pair automatically. High resolution can be obtained with small computational load.