未知噪声背景下的波达方向及载波频率三维联合估计方法

A 3D Fiducial-Referential UN-ESPRIT Algorithm Suitable for Joint Azimuth-Elevation-Carrier Estimation in Unknown Noise Environment

设定了一个具有信号子空间旋转不变性的传感器阵列以及一个任意流形的参考阵列 ,充分利用其加性噪声之间的无关性 ,提出了一种基于总的最小平方意义下的三维基准 -参考 ( 3D Fidu-cial- Referential) UN- ESPRIT算法 ,以完成在未知相关噪声背景下的二维波达角及载波频率 3D联合参量估计的任务。该方法不需要一般 UN- ESPRIT方法对于两组传感器必须具有相同阵列流形以及旋转不变性的要求 ,可以同时获得波达方向与载波频率的估计 ,且无需任何谱峰搜索过程。计算机仿真试验也证实了算法的有效性 ,并给出了较好的仿真结果。

Existing algorithms for 3D joint azimuth elevation carrier estimation are, in our opinion, not satisfactory in unknown correlative noise environment; but correlative noise is generally found in real environment. UN ESPRIT algorithm is satisfactory for 1D direction estimation in unknown noise. We propose extending the 1D UN ESPRIT idea into a 3D one; in other words, we propose a 3D Fiducial Referential UN ESPRIT (FR UN ESPRIT) algorithm that is suitable for joint azimuth elevation carrier estimation in unknown correlative noise environment. According to the proven fact that most unknown noises are spatially limited, this paper sets as fiducial array a uniform rectangular array of sensors which possess the character of rotational invariance, and sets as referential array an arbitrary array. The noise of fiducial array and that of the referential array are independent; as a result, the detection of signals in correlative noises is made easier. Even though UN ESPRIT algorithm can be extended to deal with 3D estimation, but the arrays are so complicated as to make the requirements on hardware too stringent. In contrast, fiducial and referential arrays in 3D FR UN ESPRIT algorithm are much simpler. By using one 4×4 fiducial array and one arbitrary referential array, our new 3D FR UN ESPRIT algorithm can perform much better than TLS ESPRIT algorithm [4] in unknown correlative noise environment.