Conference
Design of Dual-Band Microstrip Patch Antenna For Wireless Communication Applications Using Machine Learning Techniques
Khairool Bariyah Mohamad Kamal
Nur Izzati Mohd Ali
M. Markom
S. Suhaimi
Raja Mariatul Qibtiah Raja Jaapar
Aug 2026 · 2026 IEEE International Conference on Electrical, Electronic and Computer Engineering (ICEECE) · pp. 1-4 · 0 citations
· 16 references
Abstract
This paper presents a dual-band microstrip patch antenna (2.4/5.2 GHz) on Rogers RT-duroid 5880, optimized using Artificial Neural Network (ANN) and Random Forest (RF) models to predict antenna characteristics.. Model effectiveness is evaluated using the Coefficient of Determination (R2 Score) and Mean Absolute Error (MAE). The ANN demonstrates superior accuracy, achieving an R2 of 0.9935 and an MAE of 0.0048, significantly outperforming the RF model. The final optimized antenna achieves a 0% frequency error at the target bands with a directivity of 6.56 dBi, confirming that the ANN provides a highly efficient solution compared to conventional electromagnetic simulations.