We implement an efficient artificial neural network (ANN), for the analysis and rapid inverse design of dielectric and planar hyperbolic metamaterial based waveguides operating in the infrared wavelength (λ) telecommunication spectrum (1200nm to 1700nm), where long propagation length and high confinement of the fields are desirables. We consider waveguides made of alternate metal-dielectric layers anisotropic claddings, surrounding a homogeneous core. The main device propagation properties to be computed by the proposed ANN are the length propagation (L) and the penetration depth (dp) associated to every design (outputs of the ANN). They are function of the λ excitation, metal (nm), dielectric (nd) refractive indexes of the layers and core refractive index (nc) as physical parameters. Our approach demonstrates high behavior prediction, design accuracy and minimal modeling parameters.
|