基于旋转变换矩阵和广义共轭剩余法的信号源数检测方法

Detection of the Number of Signals Using the Transformed Rotational Matrix and General Conjugate Residue Method

确定空间信号源数目在高分辫率测向中起着至关重要的作用。提出了一个鲁棒的误检概率较低的检测信号源数目算法——旋转变换矩阵-广义共轭剩余法(VTRS-GCR),算法主要利用了旋转变换子阵方差(VTRS)的趋零性,同时使用广义共轭剩余法(GCR)来解矩阵方程得到旋转变换矩阵的估计值。算法适用于任何的子阵几何模型,同时不需要额外的参数和主观的门限设定。如果与空间平滑技术相结合,还可以用于检测相关信号。仿真结果表明,VTRS-GCR方法在较低的信噪比和较小的快拍数条件下都能表现良好的检测效果,同时还可用于色噪声背景和信号源数目较大的条件下。

Detection of the number of signals plays an important role in high resolution bearing estimation algorithms. A robust number-of-source detection method is presented with low detection error rate-VTRS-GCR,which utilizes the tendency to zero of the variance of transformed rotational submatrix （VTRS） and the general conjugate residue method （GCR） to obtain the estimate of transformed rotational matrix. This method applies to arbitrary to subarray geometries. Besides, no subjective threshold or extra parameter is required. Combined with spatial smoothing,this method can also detect coherent signals. According to the simulation, this method shows high robustness in the case of low SNR and small snapshot. In addition, it can also perform very well when the additive noise is colored noise or the number of signal is large.