A Dual-Layer Coupled Network Perspective on Shared-Aperture Transparent Antenna Arrays for Integrated Space-Air-Ground Networks

This paper presents a network-theoretic modeling and design framework for dual-band shared-aperture tightly coupled dipole arrays (TCDAs) deployed in space-aerial-terrestrial integrated networks (SATIN). The proposed system integrates a low-frequency array (0.6-2.4 GHz) and a high-frequency array (3.5-12 GHz) in a spatially co-located configuration, forming a dual-layer electromagnetic structure. A graph-based abstraction is introduced to characterize the coupled behavior across layers, with each sub-array treated as a frequency-specific network layer and interlayer coupling modeled as weighted edges. To mitigate cross-band interference and enhance vertical transparency, frequency-selective meander-line filters are embedded in the lowfrequency dipole arms, enabling effective electromagnetic decoupling without compromising bandwidth or radiation efficiency. Machine learning-assisted surrogate modeling is employed to accelerate prediction of impedance matrices and active element patterns, facilitating optimization of array excitation and systemlevel metrics such as channel capacity and gain. A 12 × 12 highfrequency and 8×8 low-frequency prototype array is fabricated and tested, demonstrating stable ±30° beam scanning, average radiation efficiencies of 92% and 88% in their respective bands, and strong alignment with simulation results. The proposed approach offers a scalable and physically grounded solution for multi-band array design in next-generation SATIN platforms.