Fast Design of Metasurface-based Microwave Absorber Using the Neuro-TF Approach

Due to its great capability in manipulating electromagnetic (EM) waves, metasurfaces have broad application prospects. Conventional approaches for metasurfaces design demand huge amounts of full-wave EM simulations to obtain the optimal geometric parameter values, leading to a CPU-intensive and time-consuming optimization process. Parametric modeling is an important part of achieving fast design optimization of microwave components. Recently, an advanced technique which combines neural networks and pole-residue-based transfer function (i.e., the neuro-TF approach) shows great potential in parametric modeling, which is so far mostly applied to microwave passive components. However, how to apply the neuro-TF method to parametric modeling and fast design of metasurfaces remains an open project. The main contribution in this paper is the development of the neuro-TF model which provides accurate and fast prediction of the EM behavior of a metasurface and thus greatly accelerate the design optimization process. The model can then be utilized to achieve fast design of metasurfaces-based components such as microwave absorbers. A frequency-selected microwave absorber is used as an example to demonstrate the superior performance of the neuro-TF approach.