Persymmetric subspace Gradient-based detectors for distributed targets in homogeneous and partially homogeneous environments

Abstract

In this paper, we consider the problem of adaptive detection for distributed targets in zero-mean Gaussian clutter with unknown persymmetric covariance matrix. The target signal is assumed to locate in a known subspace with unknown coordinates. Two persymmetric subspace detectors are designed by utilizing the Gradient criterion in homogeneous and partially homogeneous environments, respectively. Additionally, the proposed detectors are theoretically confirmed to be constant false alarm rate to the clutter covariance matrix and the power level. Moreover, numerical results show that the proposed detectors perform better than the existing competitors, especially in training-limited scenarios.