Robust CFAR detection of random signals in compound-Gaussian clutter plus thermal noise

An algorithm for detecting a random target signal against a mixture of correlated compound-Gaussian clutter and white Gaussian thermal noise is proposed. The new detection strategy is obtained by extending the generalised matched subspace detector previously derived by Gini and Farina (1999) for only compound-Gaussian clutter. Two different versions of the detection strategy are proposed and compared: the first relies on the estimation of the radar clutter texture component; the second trades-off performance with computational complexity by using the texture mean value in place of its estimate. The texture estimator mean square error is derived in closed form and analysed. Additionally, the robustness of the detector false alarm rate to changes of clutter parameters and the detection performance are numerically investigated by Monte Carlo simulation.