Related Experiment Video
Updated: Jan 11, 2026
13:44
Simulation, Fabrication and Characterization of THz Metamaterial Absorbers
Published on: December 27, 2012
15.8K
Metamaterial absorber using cascaded ring resonators and optimization through machine learning for sensing
Mohammad Reza Rakhshani1,2, Fatemeh Kazemi3, Mahdi Rashki4
1Department of Electrical Engineering, Faculty of Engineering, University of Zabol, Zabol, 9861335856, Iran. mrakhshani@uoz.ac.ir.
Scientific Reports
|November 14, 2025
Summary
This study introduces an Adaptive Neuro-Fuzzy Inference System (ANFIS) to design metamaterial absorbers (MA), significantly reducing simulation time by 90% compared to traditional methods. The ANFIS model accurately predicts absorption, enabling efficient development of MAs for applications like sensors.
Area of Science:
- Metamaterial research
- Computational electromagnetics
- Machine learning applications in physics
Background:
- Conventional metamaterial absorber (MA) design relies on computationally intensive methods like finite-difference time-domain (FDTD).
- These trial-and-error simulations are time-consuming and limit rapid design iterations.
- There is a need for efficient and accurate methods to design complex nanostructures like cascaded ring resonators.
Purpose of the Study:
- To develop a cost-effective and time-efficient design framework for narrowband metamaterial absorbers (MA).
- To utilize an Adaptive Neuro-Fuzzy Inference System (ANFIS) as an alternative to traditional FDTD simulations.
- To demonstrate the ANFIS model's capability in predicting MA absorption based on geometrical parameters.
Main Methods:
- Designed a narrowband metamaterial absorber (MA) using a Metal-Insulator-Metal (MIM) structure with cascaded ring resonators (RR).
- Employed an Adaptive Neuro-Fuzzy Inference System (ANFIS), a hybrid neural network and fuzzy logic model, for absorption prediction.
- Utilized four geometrical parameters as input variables for the ANFIS model and validated performance using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) with 5-fold cross-validation.
Main Results:
- The ANFIS model achieved high prediction precision with RMSE of 0.0811 and MAE of 0.062.
- ANFIS demonstrated approximately 90% time savings compared to the conventional FDTD method.
- The proposed MA exhibited nearly perfect absorption and achieved high sensitivity (600 nm/RIU) as a sensor.
Conclusions:
- ANFIS provides a highly accurate and significantly faster alternative for designing metamaterial absorbers.
- The proposed metamaterial absorber demonstrates excellent performance for sensing applications.
- The tunable and compact MA design holds potential for diverse applications including solar cells, sensors, and stealth technology.
