eduzhai > Applied Sciences > Engineering >

A Generative Machine Learning-Based Approach for Inverse Design of Multilayer Metasurfaces

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 12 pages

Abstract: The synthesis of a metasurface exhibiting a specific set of desiredscattering properties is a time-consuming and resource-demanding process, whichconventionally relies on many cycles of full-wave simulations. It requires anexperienced designer to choose the number of the metallic layers, the scatterershapes and dimensions, and the type and the thickness of the separatingsubstrates. Here, we propose a generative machine learning (ML)-based approachto solve this one-to-many mapping and automate the inverse design of dual- andtriple-layer metasurfaces. Using this approach, it is possible to solvemultiobjective optimization problems by synthesizing thin structures composedof potentially brand-new scatterer designs, in cases where the inter-layercoupling between the layers is non-negligible and synthesis by traditionalmethods becomes cumbersome. Various examples to provide specific magnitude andphase responses of $x$- and $y$-polarized scattering coefficients across afrequency range as well as mask-based responses for different metasurfaceapplications are presented to verify the practicality of the proposed method.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×