Metamaterial Lens Design for Space Optics using Physics-Informed Neural
Networks
Dylan Everingham
(COSSE student, double degree with
TU
Berlin)
Site of the project: TU Delft
Supervisor TU Delft: Matthias Moller and Aurele Adam (Optics group)
start of the project: October 2021
In March 2022 the
Interim Thesis
has appeared
and a
presentation
has been given.
Summary of the master project:
Artificial composite materials which contain lattice nanostructure and
exhibit electromagnetic response in the terahertz domain, which we call
terahertz metamaterials, hold great potential for flat optics which can
replace bulky lens arrays. However, traditional numerical inverse
scattering models for predicting the parameters of a desired lens become
ill-posed in the presence of strong multiple light scattering, such as in
the case of a complex metamaterial lens lattice topology. I am
investigating the use of physics-informed neural networks (PINNs) as an
alternate method for solving this inverse scattering problem, with the goal
of predicting an effective lens design for a specific application in space
optics.
Contact information:
Kees
Vuik
Back to the
home page
or the
Master students page of Kees Vuik