A Novel Compact Design of Multi-Band Frequency Selective Surface for 5G Applications using a CNN-GNN Hybrid Network Optimization

Abstract

In this paper, a new compact design of multi-band Frequency Selective Surface (FSS) optimized using a hybrid Convolutional Neural Network-Graph Neural Network (CNN-GNN) framework for advanced 5G communication systems is proposed. The suggested FSS operates in five different frequency bands, i.e., 4.0788 GHz, 17.080 GHz, 21.559 GHz, 27.928 GHz and 29.870 GHz with bandwidths of 2.686 GHz and 1.849 GHz in important bands. Unlike the usual designs with only one or two bands with limited bandwidths, the proposed FSS provides significant multi-band capabilities while maintaining excellent angular stability up to 80° for both Transverse Electric (TE) and Transverse Magnetic (TM) polarizations. In addition, the design with the compact unit cell size of 0.0814λ0×0.0814λ0 is suitable for integration into the contemporary 5G devices, including the wearable and automotive systems. The FSS displays effective cross-polarization suppression, directional radiation patterns, and minimal backward interference, with enhanced antenna gain and dependable signal filtering at varied incident angles, through rigorous simulation and experimental validation. The implementation of a CNNGNN hybrid optimization framework to improve the FSS shape is one of the main contributions of this work, combining the characteristics of localized feature extraction and global dependency modeling. The hybrid technique results in a decrease of return loss and better selectivity in frequency than traditional designs and better performance metrics. Experimental results are in good agreement with simulations, confirming the robustness of the design and its practical applicability. The 
suggested FSS surpasses the existing solutions with respect to multi-band operation, polarization stability and bandwidth. It 
is a better choice for several 5G applications such as MIMO systems and mmWave technologies. The study solves important 
gaps in the state-of-the-art via a compromise between compactness, efficiency and advanced optimization, opening the way 
for scalable and high-performance 5G infrastructures.