Clutter Edges Detection Algorithms for Structured Clutter Covariance Matrices

This letter deals with the problem of clutter edge detection and localization in training data. To this end, is formulated as a binary hypothesis test assuming that ranks covariance matrix are known, adaptive architectures designed based on generalized likelihood ratio to decide whether data within sliding window contains homogeneous set or two heterogeneous subsets. In design stage, we utilize four different structures (i.e., Hermitian, persymmetric, symmetric, centrosymmetric) exploit priori information. Then, for case unknown ranks, extended by devising preliminary estimation stage resorting model order selection rules. Numerical examples both synthetic real highlight proposed solutions possess superior performance respect competitors do not use any