CFAR property and robustness of the lowrank adaptive normalized matched filters detectors in low rank compound gaussian context

Abstract

In the context of a heterogeneous disturbance with a Low Rank (LR) structure (referred to as clutter), one may use the LR approximation for detection process. Indeed, in such context, adaptive LR schemes have been shown to require less secondary data to reach equivalent performances as classical ones. The LR approximation consists of canceling the clutter rather than whitening the whole noise. The main problem is then the estimation of the clutter subspace instead of the noise covariance matrix itself. Maximum Likelihood estimators (MLE), under different hypothesis [1][2][3], of the clutter subspace have been recently proposed for a noise composed of a LR Compound Gaussian (CG) clutter plus a white Gaussian Noise (WGN). This paper focuses on the numerical analysis of performances of the LR Adaptive Normalized Matched Filter (LR-ANMF) detectors build from these different clutter subspace estimators. Numerical simulations and a real data set illustrate their CFAR property with respect to heterogeneity and robustness to outliers.