基于深度学习的线性偏振均匀聚焦的超构透镜逆向设计  

Inverse Design of Meta-Lenses for Linearly Polarized Focusing with Uniform Field Distribution via Deep Learning

作　　者：迟浩翔 Haoxiang Chi(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海

出　　处：《建模与仿真》2024年第3期3635-3642,共8页Modeling and Simulation

摘　　要：为了实现线偏振的多焦点的能量均匀调控,提出了一种基于深度学习和逆向算法相结合的设计方法实现的多焦能量均匀调控的全介质超构透镜。利用全连接神经网络预测不同尺寸的单元结构,结合时域有限差分方法(FDTD)来实现逆向设计并进行数值仿真验证。该设计方法表明,通过对两个正交偏振态分别进行振幅和相位分布的逆向设计,实现了在x偏振态下的横向三个能量均匀的焦点,在y偏振态下的纵向三个能量均匀的焦点。这种独特的设计超表面的方式为开发太赫兹波段的高精度功能器件提供了新的路径,有望应用于成像、探测、传感。In order to realize the uniform energy control of linearly polarized multi-focus,a novel design method based on deep learning and inverse algorithm is proposed to achieve the uniform energy control of multi-focus all-dielectric metaslens.The fully connected neural network is used to predict the unit-cell structures of different sizes,and the Finite Difference Time Domain(FDTD)method is used to obtain the reverse design and the numerical simulation is verified.Our design methodology demonstrates the inverse design of amplitude and phase distributions for two orthogonal polarization states,resulting in three uniformly distributed focal points in the transverse direction for x-polarized state and three uniformly distributed focal points in the longitudinal direction for y-polarized state.This unique approach to designing metasurfaces offers a promising avenue for the development of high-precision devices in the terahertz frequency band,with potential applications in sensing,imaging and detection.

关 键 词：超表面 深度学习 焦点 均匀能量 逆向设计 

分 类 号：O43[机械工程—光学工程]