Statistical Interpretation of a Data Adaptive Clutter Subspace Estimation Algorithm

Numerical performances of low rank stap based on different heterogeneous clutter subspace estimators

“…However, [1] [2] show that, in the considered framework of LR-Compound Gaussian plus white Gaussian noise, the clutter subspace projector MLE is derived from a matrix that is not an estimate of the CM. This intermediary matrix is the SCM of the data scaled by a factor that gives more significance to samples that have strong power into the subspace of interest.…”

“…Unlike the classical filter which is based on the Covariance Matrix (CM) of the noise, the LR filter is based on the clutter subspace projector, which is usually derived from a Singular Value Decomposition (SVD) of a noise CM estimate. Regarding to the considered model of LR-CG plus WGN, recent results are providing both direct estimators of the clutter subspace [1][2] and an exact MLE of the noise CM [3]. To promote the use of these new estimation methods, this paper proposes to apply them to realistic STAP simulations.…”

“…Regarding to the considered model of LR-CG plus WGN, recent results are providing both direct estimators of the clutter subspace [1][2] and an exact MLE of the noise CM [3]. In [1] is proposed an estimator of the clutter subspace under specific hypotheses: the CM of the low-rank CG clutter is assumed to have identical eigenvalues, and the Probability Density Function (PDF) of the texture is assumed to be known.…”

“…In [1] is proposed an estimator of the clutter subspace under specific hypotheses: the CM of the low-rank CG clutter is assumed to have identical eigenvalues, and the Probability Density Function (PDF) of the texture is assumed to be known. The assumption of known texture PDF has been relaxed in [2] by considering texture as an unknown deterministic parameter.…”

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

“…A currently an active topic of research focuses on regularization of the algorithms to compute these estimators in under sampled configurations [9] [10]. Nevertheless, with regard to the considered model of LR-CG plus WGN, recent results are providing direct CSP estimators [1][2] [3]. These estimators are derived from a intermediate matrix that is not necessarily an estimate of the CM and can be computed when K > R.…”