Loading...

Design of fresnel-region millimeter-wave metasurface beam shaper using deep learning

Koohi Ghamsari, M. H ; Sharif University of Technology | 2024

34 Viewed
  1. Type of Document: Article
  2. DOI: 10.1109/ICEE63041.2024.10668240
  3. Publisher: IEEE , 2024
  4. Abstract:
  5. In this paper, we present a new approach for designing a metasurface structure that leverages deep learning techniques for beam shaping a millimeter-wave source in the Fresnel region. The selection of unit cell geometry is crucial to attain accurate control over the phasefront of the incident wave. We show that our selection of unit cells enables the metasurface to produce the desired beam intensity on the target plane. We generate 61,000 datasets using analytical solutions of the beam shaping problem and train our neural network with an average loss of lower than 2%. The efficiency of our suggested method is evaluated by designing the phase profile of the metasurface to generate flat-top irradiation at a specific distance from the metasurface. The results demonstrate the feasibility of the proposed deep learning approach for designing beam shaper metasurfaces. © 2024 IEEE
  6. Keywords:
  7. Beam shaping ; Deep-learning ; Flat-top beam shaper ; Fresnel region ; Metasurface ; Beam forming networks ; Beamforming ; Hadrons
  8. Source: Iranian Conference on Electrical Engineering, ICEE ; Issue 2024 , 2024 ; 21647054 (ISSN)
  9. URL: https://ieeexplore.ieee.org/document/10668240