EUVIMEDEUVIMED entdecken
Crossref

Continuous Training Sample Weighting via Constrained Optimization for Robust STAP in Heterogeneous Clutter

Ciyuan Liu, Tong Wang

IEEE Signal Processing Letters · 2026 · 33 · 3147-3151

DOI: 10.1109/lsp.2026.3718340 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Ciyuan Liu und 1 weiteren Autor:innen. Die Arbeit nennt 22 Literaturangaben. Gefördert von: National Key R&D Program of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Space-time adaptive processing (STAP) relies on reliable clutter-plus-noise covariance matrix (CNCM) estimation from secondary training samples. In airborne radar, however, target-like outliers and nonhomogeneous clutter often make the training data inconsistent with the cell under test (CUT), which can seriously reduce clutter suppression performance. To address this problem, this letter develops a continuous sample weighting method for robust STAP covariance estimation. Each secondary sample is assigned a soft contribution coefficient rather than being simply retained or removed. The proposed method designs a weighting criterion that assesses the consistency between the CUT and the weighted covariance using a generalized power measure inspired by the Mahalanobis distance, suppresses abnormal samples using logarithmic scores normalized by median and median absolute deviation (MAD) statistics, and uses a Kullback-Leibler (KL)-inspired term to stabilize the weights around the reliability prior. The continuous weights are obtained by solving a convex optimization problem on the simplex with an exponentiated gradient method based on mirror descent. Numerical results in homogeneous, outlier-contaminated, and nonhomogeneous clutter scenes verify the effectiveness of the proposed method.

Förderung

  • National Key R&D Program of China

Zitieren

APA

Liu, C., & Wang, T. (2026). Continuous Training Sample Weighting via Constrained Optimization for Robust STAP in Heterogeneous Clutter. IEEE Signal Processing Letters, 33, 3147-3151. https://doi.org/10.1109/lsp.2026.3718340

Vancouver

Liu C, Wang T. Continuous Training Sample Weighting via Constrained Optimization for Robust STAP in Heterogeneous Clutter. IEEE Signal Processing Letters. 2026;33:3147-3151. doi:10.1109/lsp.2026.3718340

BibTeX
@article{Liu2026,
  author = {Liu, Ciyuan and Wang, Tong},
  title = {Continuous Training Sample Weighting via Constrained Optimization for Robust STAP in Heterogeneous Clutter},
  journal = {IEEE Signal Processing Letters},
  year = {2026},
  volume = {33},
  pages = {3147--3151},
  doi = {10.1109/lsp.2026.3718340}
}
Kostenlos starten →