Reduced-cost two-level surrogate antenna modeling using domain confinement and response features.

Area of Science:

• Engineering
• Computational Electromagnetics
• Antenna Design

Background:

• Electromagnetic (EM) simulation is crucial for antenna design but computationally expensive.
• High computational costs limit the feasibility of iterative design tasks.
• Surrogate models (metamodels) offer a solution by replacing EM simulations with faster approximations.

Purpose of the Study:

• To develop a novel modeling technique for efficient antenna design.
• To address the curse of dimensionality in surrogate modeling for antennas.
• To reduce the reliance on extensive electromagnetic simulations.

Main Methods:

• Incorporation of response feature technology into a constrained modeling framework.
• Development of reference-design-free constrained modeling techniques.
• Creation of metamodels restricted to high-quality design parameter spaces.

Main Results:

• Accurate surrogate models can be generated using exceptionally small training datasets.
• The proposed technique significantly reduces computational overhead in antenna design.
• The methodology is applicable to antennas with consistent input characteristic shapes.

Conclusions:

• The novel technique offers a computationally efficient approach to antenna design.
• This method enables the creation of accurate metamodels with minimal data.
• The approach has potential applications in other fields with costly simulation models.