Radar detection and range estimation of a point-like target in non-Gaussian, possibly correlated, noise

In this paper, we propose an adaptive algorithm to detect and localize a point-like target in heterogeneous environments. At the design stage we use the two-step GLRT-based design procedure: we first introduce the GLRT for known structure of the noise covariance matrix and unknown noise levels, possibly different from one cell to another; then, we make the detector fully adaptive by replacing the unknown structure with a recursive estimate based on a set of secondary data. The performance assessment, conducted assuming noise modeled in terms of a compound-Gaussian process, shows that the newly-introduced scheme possesses good detection and localization capabilities also in comparison to its natural competitor.