EUVIMEDEUVIMED entdecken
Crossref

Heuristic Physics-Aware Neural Network Method for the Synthesis of Sparse Antenna Arrays

Yuhan Fan, Haotian Shi, Ya Fei Wu, Yongxin Guo

IEEE Transactions on Aerospace and Electronic Systems · 2026 · 62 · 12625-12637

DOI: 10.1109/taes.2026.3704820 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Aerospace and Electronic Systems, 2026. Verfasst von Yuhan Fan und 3 weiteren Autor:innen. Die Arbeit nennt 32 Literaturangaben. Gefördert von: National Research Foundation, in part by the Singapore, Future Communications Research and Development Program, Start-Up Grant for Professor.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This article proposes a dataset-free heuristic physics-aware neural network (PANN) method for optimizing planar isophoric sparse array layouts, aiming to minimize sidelobe levels under multiple physical constraints, including minimum element spacing, scanning range, aperture size, and the number of elements. First, a customized iterative grid-mutation procedure guided by marginal-effect analysis is employed to explore a quasi-optimal heuristic layout. Subsequently, an iterative learning framework is developed to evolve the heuristic layout within a continuous neighboring solution space. The framework integrates a neural network for inverse layout design with an analytical module for layout-to-performance mapping, forming a dataset-free PANN. Benefiting from the heuristic mechanism, the embedded collective mutation strategy, and gradient-based mutation operations, the proposed PANN exhibits strong exploitation capability and enhanced physical interpretability. Numerical results demonstrate the scalability of the proposed method under multiple physical constraints, ranging from small-scale arrays ($N 1000$), as well as its superior performance compared with state-of-the-art methods. For example, in small-scale array synthesis, the proposed method outperforms existing approaches by at least 3.1 dB. Moreover, even with a further 32% reduction in the number of elements, it still achieves an additional improvement of 0.40 dB.

Förderung

  • National Research Foundation, in part by the Singapore
  • Future Communications Research and Development Program
  • Start-Up Grant for Professor
  • City University of Hong Kong

Zitieren

APA

Fan, Y., Shi, H., Wu, Y. F., & Guo, Y. (2026). Heuristic Physics-Aware Neural Network Method for the Synthesis of Sparse Antenna Arrays. IEEE Transactions on Aerospace and Electronic Systems, 62, 12625-12637. https://doi.org/10.1109/taes.2026.3704820

Vancouver

Fan Y, Shi H, Wu YF, Guo Y. Heuristic Physics-Aware Neural Network Method for the Synthesis of Sparse Antenna Arrays. IEEE Transactions on Aerospace and Electronic Systems. 2026;62:12625-12637. doi:10.1109/taes.2026.3704820

BibTeX
@article{Fan2026,
  author = {Fan, Yuhan and Shi, Haotian and Wu, Ya Fei and Guo, Yongxin},
  title = {Heuristic Physics-Aware Neural Network Method for the Synthesis of Sparse Antenna Arrays},
  journal = {IEEE Transactions on Aerospace and Electronic Systems},
  year = {2026},
  volume = {62},
  pages = {12625--12637},
  doi = {10.1109/taes.2026.3704820}
}
Kostenlos starten →