Fusion Neural Networks for High-Precision Design and Ultrawideband Shielding in Frequency Selective Surfaces
- 16 December 2024
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Components, Packaging and Manufacturing Technology
- Vol. 15 (4) , 810-820
Abstract
This research advances a multi-input, multi-output regression approach for continually optimizing the physical design of frequency-selective surfaces (FSS) to enhance shielding effectiveness over an ultra-wide bandwidth. The classical models of neural networks are frequently used for such regression tasks due to the capacity to extract features from average datasets; however, they struggle with complex design parameters and broad responses. To accomplish this, a hybrid neural network technique is developed, which combines convolutional neural networks (CNN) for deep feature capture with LSTM networks and an attention mechanism for processing. This method makes efficient use of linear information in design dimensions while also consistently computing S-parameters. The FSS design employs a double square loop that resonates at two frequencies, resulting in a wideband response, by adding four stubs between the loops it becomes an ultra-wideband response. The equivalent lumped circuit (ELC) model is applied for estimating capacitance (C) and inductance (L), which are then transformed to ABCD and S-parameters. The model obtained a r 2 value of 99.3% and a mean squared error (MSE) of 1 × 10 –3 after optimizing the design parameters in 124.25 seconds. The design, implemented on Rogers 5880LZ with a thickness of 1.27 mm and unit cell dimensions of 11×11 mm², is stable across varied incidence angles and polarization insensitive, allowing for downsizing. Comparative research demonstrates that the CNN-LSTM hybrid model surpasses traditional techniques, illustrating higher shielding effectiveness.Keywords
This publication has 24 references indexed in Scilit:
- Reaching the efficiency limit of arbitrary polarization transformation with non-orthogonal metasurfacesNature Communications, 2024
- Design Methodology of Dual-Polarized Angle-Selective Surface Based on Three-Layer Frequency-Selective SurfacesIEEE Transactions on Antennas and Propagation, 2023
- Miniaturized Near All-Angle Stable 2.5-D FSS for Dual-Band Shielding in Transmission Mode and Multiband Absorption in Reflection ModeIEEE Transactions on Antennas and Propagation, 2023
- Using an Equivalent-Circuit Model to Design Ultra-Wide Band-Stop Frequency-Selective Surface for 5G mm-Wave ApplicationsIEEE Open Journal of Antennas and Propagation, 2022
- Adversarial-Network Regularized Inverse Design of Frequency-Selective Surface With Frequency-Temporal Deep LearningIEEE Transactions on Antennas and Propagation, 2022
- FSS Derived Using a New Equivalent Circuit Model Backed Deep Neural NetworkIEEE Antennas and Wireless Propagation Letters, 2021
- Fourier Subspace-Based Deep Learning Method for Inverse Design of Frequency Selective SurfaceIEEE Transactions on Antennas and Propagation, 2021
- Compact Ultra-Wide Band Frequency Selective Surface With High SelectivityIEEE Transactions on Antennas and Propagation, 2020
- Automatic design of inductive FSSs using the genetic algorithm and the MoM/BI-RME analysisIEEE Antennas and Wireless Propagation Letters, 2002
- A new genetic algorithm-based frequency selective surface design for dual frequency applications1999