Response of sample covariance based MVDR beamformer to imperfect look and inhomogeneities

The sample covariance based (SCB) minimum variance distortionless response (MVDR) beamformer has desirable properties under ideal homogeneous data conditions with perfect look. Such properties include maximum-likelihood optimality yielding an unbiased asymptotically efficient estimate of the signal parameter. When target array response mismatch exists and inhomogeneities are present, however, its performance can degrade considerably. Exact means and variances are given for the SCB MVDR beamformer output accounting simultaneously for finite sample support, imperfect look, and inhomogeneities demonstrating biased signal estimates and potentially extreme vulnuerability to undernulled interference and surprise jammers.