阵列测向中基于投影分解的信源数目后判断

Posteriori source number decision based on subspace projection in sensor array processing

信源数目判断是阵列测向中的重要课题，目前常用的方法一般都不利用信源的入射方向估计信息，称为信源数目的“前”判断方法。由于使用的信息较少，这类方法很难给出信源数目的准确估计。当信源数目估计过少时，将影响测向效果，而过多时则无法去掉虚假信号。为解决该问题，提出了基于投影分解的信源数目“后”判断方法。分析证明了该方法与ＭＵＳＩＣ、极大似然算法配合使用的问题，并进行了计算机模拟。结果表明，该方法有较强的适用性且判断效果较好。

This paper deals with the problem of source number decision in sensor array processing. Most of the existing source number decision methods do not use the direction of arrival (DOA) estimation results, and can be called a priori decision method. These methods can not give very accurate source number decision because of the lack of DOA information. The sensor array processing algorithms would suffer from a poor performance when the estimated source number is less, and there would exist un removable false results when it is more. A posteriori source number decision method based on the subspace projection to solve the above defects is proposed. Analysis and simulation show that this method can easily match the widely used DOA estimation method such as MUSIC, maximum likelihood estimation, and subspace fitting.