A physics-informed deep neural network for low sidelobe cosecant-squared pattern synthesis

Tarek Sallam, Qun Wang and Ahmed M. Attiya

Journal of Electromagnetic Waves and Applications, 2025, vol. 39, issue 16, 1985-1996

Abstract: In this paper, we present a novel scheme to train a deep neural network (DNN) for a cosecant-squared (csc2) radiation pattern synthesis speedily and accurately without relying on numerical optimization methods. Our approach is to train a DNN using a physics-informed loss function that minimizes the deviation between the desired shaped beam pattern and the actual one. The physics-informed deep neural network (PIDNN) makes array synthesis effective because it can efficiently achieve two desired features, namely, low sidelobe level (SLL) and small deviation (ripples) in the shaped beam region. To obtain a desired csc2 pattern with the SLL constrained, PIDNN optimizes the excitation amplitude and phase weights of the array elements. To illustrate the effectiveness of the proposed method, the beam pattern with specified characteristics is obtained for the same array by using genetic algorithm (GA) and particle swarm optimization (PSO).

Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://hdl.handle.net/10.1080/09205071.2025.2532813 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:tewaxx:v:39:y:2025:i:16:p:1985-1996

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/tewa20

DOI: 10.1080/09205071.2025.2532813

Access Statistics for this article

Journal of Electromagnetic Waves and Applications is currently edited by Mohamad Abou El-Nasr and Pankaj Kumar Choudhury

More articles in Journal of Electromagnetic Waves and Applications from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().