Medium metasurface reverse design algorithm utilizing cascaded deep neural network
A deep neural network, reverse design technology, applied in the field of reverse design algorithms of medium metasurfaces, can solve problems such as inability to provide solutions
Inactive Publication Date: 2021-01-12
NANKAI UNIV
View PDF0 Cites 1 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Some current traditional methods such as particle swarm algorithm, evolutionary algorithm and reverse design of simulated annealing algorithm cannot provide an effective solution
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0039] Select the working band as the visible light band 450-850nm, in this band the material of the dielectric nanopillar is GaN, and the substrate material is Al 2 o 3 , the material combination has a high refractive index and an absorptivity close to zero in the working wavelength range. The period of the response unit is determined to be 400nm by applying Quinstein's sampling law and the constraints of processing conditions. Four cuboids were randomly generated within the half-side period and mapped to the other side in a mirror-symmetrical manner. The length and width resolution of the generated nanopillars was set to 10 nm, and the height was 500, 600, or 700 nm. A 20nm blank was set at the edge to reduce the coupling between the response units, and then the structural features of each response unit were recorded in the form of a gray scale image (gray scale represents the height). The electromagnetic response of the structure is calculated by the method of finite diff...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a medium metasurface reverse design algorithm using a cascaded deep neural network. The method comprises the following steps: step 1, selecting a working waveband and a super-surface response unit dielectric material, randomly generating a plurality of medium super-surface response units, and calculating electromagnetic response of each response unit by using a time domainfinite difference method to form a one-to-one correspondence database; step 2, constructing a cascaded deep neural network; and 3, training a forward network of the cascaded deep neural network by taking the picture containing the structure information as input and taking the electromagnetic response as output. And step 4, training an auto-encoder for dimensionality reduction and feature extraction of a picture containing structural information. And step 5, testing the performance and generalization ability of the cascaded deep neural network. The invention can be applied to rapid prediction and targeted reverse design of the transmission spectrum of the medium metasurface response unit, and the time cost and computing power of design are greatly saved.
Description
technical field [0001] The invention relates to the technical fields of micro-nano optics and machine learning, in particular to a medium metasurface reverse design algorithm using a cascaded deep neural network. Background technique [0002] A metasurface is a two-dimensional device composed of subwavelength structures that can achieve arbitrary modulation of the amplitude, phase, and polarization of incident light. In recent years, metasurfaces have been widely used in holography, metalens and special beam generators, etc. The current metasurface design methods rely on methods such as finite-difference time domain or finite element methods. Although some commercial software has been established for the calculation of the electromagnetic response of the metasurface response unit, the parameter scanning of the response unit requires a lot of time and computing power. In addition, the design and extent of metasurfaces require specific amplitude and phase distributions, so t...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More IPC IPC(8): G06F17/10G06N3/04G06N3/08
CPCG06F17/10G06N3/08G06N3/045
Inventor 匡登峰孔维超
Owner NANKAI UNIV