Enabling Intelligent Metasurfaces for Semi-known Input

作者：Pujing Lin, Chao Qian, Jie Zhang, Jieting Chen, Xiaoyue Zhu, Zhedong Wang, Jiangtao Huangfu, Hongsheng Chen · 发表于：Electromagnetic waves · 年份：2023 · DOI：10.2528/pier23090201 · 被引用次数：79 · 研究领域：Advanced Antenna and Metasurface Technologies、Opportunistic and Delay-Tolerant Networks

Compelling evidence suggests that the interaction between electromagnetic metasurfaces and deep learning gives rise to the proliferation of intelligent metasurfaces in the past decade.In general, deep learning offers a transformative force to reform the design and working style of metasurfaces.Most of the inverse-design literature announces that, given a user-defined input, pre-trained deep learning models can quickly output the metasurface candidates with high fidelity.However, they largely ignore an important fact, that is, the practical input is always semi-known.In this work, we introduce a generation-elimination network that is robust to semi-known input and information pollution.The network is composed of a generative network to generate a number of possible answers and then a discriminative network to eliminate suboptimal answers.We benchmark the feasibility via two scenes, the on-demand metasurface design of the reflection spectra and the far-field pattern.In the microwave experiment, we fabricated and measured the reconfigurable metasurfaces to automatically meet the semi-known beam steering requirement that widely exist in wireless communication.Our work for the first time answers the question of how to cope with semi-known input, which is ubiquitous in a panoply of real-world applications, such as imaging, sensing, and communication across noisy environment.