Multifrequency Near-Field Focusing Through Infinitesimal Dipole Models and Neural Networks
Year: 2025
Type: article
Abstract: We introduce a multi-frequencies near-field focusing (NFF) technique based on the infinitesimal dipole model (IDM) and enhanced by neural networks. By deriving the IDM from the active element pattern (AEP), our method accurately estimates the near-field behavior of array systems, even when affected by electromagnetic mutual coupling. For single-frequency NFF, we employ semi-definite convex optimization, while a neural network predicts optimal array excitation weights for multiple frequencies. This approach enables rapid NFF at multiple frequencies. When applied to a half-wavelength dipole array, our method demonstrates high-precision NFF within milliseconds and achieves significant computational efficiency. This makes it particularly well-suited for multi-frequencies near-field communication applications anticipated to be pivotal in 6 G wireless technologies.
Authors: Jiarun Hu, Said Mikki
Institutions: University of Illinois Urbana-Champaign, Zhejiang University
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $2,645
OpenAlex ID: W4409473743