Random Matrix-Optimized High-Dimensional MVDR Beamforming

摘要

A new approach to minimum variance distortionless response (MVDR) beamforming is proposed under the assumption of simultaneously large numbers of array sensors and observations. The key to our method is the design of an inverse covariance estimator which is appropriately optimized for the MVDR application. This is obtained by exploiting spectral properties of spiked covariance models in random matrix theory. Our proposed solution is simple to implement and is shown to yield performance improvements over competing approaches.