A new Bayesian parametric glrt for multichannel adaptive signal detection

This paper considers the problem of knowledge-aided space-time adaptive processing (KA-STAP) combined with a parametric technique. Specifically, by modeling the disturbance as a multichannel autoregressive (AR) process, we introduce a stochastic signal model in which the spatial covariance matrix of the disturbance is assumed to be random, with some prior distribution. Incorporating the a priori knowledge into the detector design stage, a Bayesian version of the parametric generalized likelihood ratio test (PGLRT) is derived based upon both test and training signals. The resulting detector, referred to as the KA-BPGLRT, can handle the challenging case where no training signals can be employed.