虚拟阵列的四阶MUSIC算法研究

Research on fourth-order MUSIC algorithm for virtual arrays

研究一种基于四阶累积量的虚拟阵列扩展技术,以扩大阵列的处理孔径和空间自由度,提高方位估计性能.由于累积量域导向矢量的冗余项可等效为特定位置处假想阵元(虚拟阵元)的响应,因而利用高阶累积量可虚拟扩展阵列孔径,利用四阶累积量与二阶统计量的转换关系,对具有等效阵元互相关的累积量作合并平均处理,可得到虚拟扩展阵列的协方差矩阵,对虚拟协方差矩阵采用MUSIC算法作方位估计.数值分析和湖试处理结果表明,虚拟阵MUSIC算法能有效提高分辨率,减小方位估计方差,并能提高空间有色高斯噪声下的性能和稳健性.

A virtual array expansion technique based on the fourth-order （FO） cumulant was studied, whereby the array processing aperture and the spatial degrees of freedom were expanded. As a result, bearing estimation performance was improved. Since each of the redundant terms of the steering vector in the cumulant domain was seen to be equivalent to the response of an assumed element at a specific position, the virtual aperture could be virtually expanded using higher-order cumulants. The virtual array covariance matrix was estimated by averaging all the FO cumulants with the same equivalent inter-element cross correlations and utilizing the fact that the second order statistics were derived from the FO cumulants. The incoming signal direction was found with the MUSIC algorithm. Numerical analysis and results from a lake trial show that the resolution is increased and estimation variance is reduced by the virtual array MUSIC algorithm, and it is robust with spatial colored Gaussian noise.