AUTHOR=Lin Hai , Tian Yuze , Hou Junjie , Xu Weilin , Shi Xinyang , Tang Rongxin 

TITLE=Fussy Inverse Design of Metamaterial Absorbers Assisted by a Generative Adversarial Network

JOURNAL=Frontiers in Materials

VOLUME=Volume 9 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/materials/articles/10.3389/fmats.2022.926094

DOI=10.3389/fmats.2022.926094

ISSN=2296-8016

ABSTRACT=The increasing demands for metasurfaces have led researchers to seek for effective inverse design methods, which are counting on the developments in optimization theory and deep learning techniques. Early approaches of the inverse design based on deep learning established unique mapping between device's geometry parameters and its designated EM characteristics. However, the generated solution based on traditional inverse design method may not be applicable due to practical fabrication conditions, the designers sometimes want to choose the most practical one from multiple schemes which can all meet the requirement of given EM indicators. A fuzzy inverse design method are quite demanded. In this paper, we proposed an fuzzy inverse design method for metamaterial absorbers based on Generative Adversarial Network (GAN). As a data-driven method, a self-build data sets are constructed and trained by GAN which contain absorber's design parameters and their corresponding spectral response. After the training process finished, it can generate multiple possible schemes which can satisfy the customized absorptivity and frequency bands for absorbers. The parameters generated by this model include structure sizes and impedance values, which indicates that has the ability to learn a variety of features. The effectiveness and robustness of the proposed method has been verified by several examples for the design of both narrowband and broadband metamaterial absorber. Our work proves a feasibility of using deep learning method to break the limits of one-to-one mapping for traditional inverse design method. This method may have profound usage for more complex EM device design problems in the future.