Detection of multidimensional Gaussian random signals in compound-Gaussian clutter plus thermal noise

In this work we derive two strategies for detecting a Gaussian distributed target signal, known to be confined to a given vector subspace, against a mixture of correlated compound-Gaussian clutter and white Gaussian thermal noise. Both the detectors are obtained as a generalization of the matched subspace filter approach and rely on the estimation of the local clutter power, the texture in the technical jargon, in the cell under test. Two different texture estimators are proposed and some preliminary numerical results are reported.