Efficient multi-objective optimization for switchable digital coding metasurface absorber empowered by machine...

Area of Science:

• Metamaterials and Nanophotonics
• Applied Electromagnetics
• Machine Learning Applications

Background:

• Metasurface absorbers are crucial for wireless security, radar communication, and biosensing.
• Traditional design methods are time-consuming and rely on experience.
• Existing machine learning approaches for metasurface absorbers are often single-objective and suboptimal.

Purpose of the Study:

• To develop a universal and efficient multi-objective optimization framework for designing multi-functional metasurface absorbers.
• To overcome the limitations of traditional and existing machine learning-based design methods.
• To expedite the design process for complex metasurface devices.

Main Methods:

• Utilized the Non-dominated Sorting Genetic Algorithm II (NSGA-II), a machine learning algorithm.
• Designed a four-channel dual-layer multi-functional digital coding metasurface absorber.
• Incorporated nine programmable encoding modes with four digital logic switches for seamless transitions.

Main Results:

• The framework efficiently handles homogeneous and heterogeneous multi-objective optimization problems.
• Achieved rapid convergence within 10 generations, outperforming conventional genetic algorithms.
• Optimized device demonstrates reduced FWHM, excellent impedance matching, simple fabrication, angular tolerance, and polarization insensitivity.

Conclusions:

• The proposed framework offers a significant advancement in designing multi-functional metasurface absorbers.
• The optimized device exhibits superior performance characteristics across multiple operating modes.
• This approach holds potential for designing achromatic metalenses, sensors, and detectors.