Open AccessClutter nulling space‐time adaptive processing algorithm based on sparse representation for airborne radar
Author(s) -
Zetao Wang,
Yongliang Wang,
Fei Gao,
Keqing Duan
Publication year - 2017
Publication title -
iet radar, sonar and navigation
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.489
H-Index - 82
eISSN - 1751-8792
pISSN - 1751-8784
DOI - 10.1049/iet-rsn.2016.0118
Subject(s) - clutter , subspace topology , algorithm , computer science , noise (video) , radar , constant false alarm rate , representation (politics) , rank (graph theory) , set (abstract data type) , sparse approximation , space time adaptive processing , artificial intelligence , mathematics , continuous wave radar , radar imaging , telecommunications , combinatorics , politics , political science , law , image (mathematics) , programming language
Traditional clutter subspace estimation algorithms try to solve the problem by estimating the clutter subspace eigenvectors from training samples, which are inevitably contaminated by noise. Actually, this can never be the best estimate, especially when the size of the training set is small. Furthermore, the omnipresent noise precludes the seeking of a pure clutter subspace. To cope with above drawbacks, the authors appeal to the sparse representation/recovery technique. From a mathematical viewpoint, the pure clutter subspace is spanned by the space‐time steering vectors corresponding to the clutter component, and it can be constructed by a suitable set of space‐time steering vectors selected from an overcomplete space‐time steering dictionary. A criterion for selecting these space‐time steering vectors is devised. Moreover, a clutter nulling type space‐time adaptive processing algorithm is derived based on the proposed criterion. The resulting algorithm only needs the knowledge of the noise power rather than the clutter rank, which is quite troublesome to estimate in practice. Numerical results with both simulated and the Mountain‐Top data demonstrate that the proposed algorithm has superior clutter suppression performance even with limited training samples.