Persymmetric Adaptive Detection of Distributed Targets With Unknown Steering Vectors

In this paper, we consider the distributed target detection problem with unknown signal signatures in Gaussian noise with unknown covariance matrix. Two adaptive detectors are proposed by using the persymmetry of the noise covariance matrix. We derive analytical expressions for the probabilities of false alarm of the proposed detectors, which indicate their constant false alarm rate properties against the noise covariance matrix. All the theoretical expressions are confirmed by Monte Carlo simulations. Numerical examples demonstrate that the proposed detectors have better detection performance than their counterparts, especially in the case of limited training data.