Design and Optimization of a Dual-Band F-Shaped Monopole Antenna for 2.45/5.8 GHz
ISM Medical Applications Using ANN and PSO
Rania Ibtissam Ben Melouka ,
Yamina Tighilt ,
Chemseddine Zebiri ,
Issa Tamer Elfergani ,
Rami Zegadi ,
Mohamed Lamine Bouknia ,
Abdelhak Ferhat Hamida ,
Merih Palandoken ,
Fatih Özkan Alkurt ,
Muharrem Karaaslan ,
Nail Alaoui and
Jonathan Rodriguez
The article presents the geometry and optimization of a small-sized dual-band F-shaped monopole antenna for medical applications in the 2.45/5.8 GHz ISM bands. The antenna is printed on a partial ground FR4 substrate, and its geometrical parameters are optimized by a hybrid procedure based on Artificial Neural Networks (ANNs) and Particle Swarm Optimization (PSO).Two ANN architectures, Feedforward Backpropagation Neural Network (FFBPN) and Generalized Regression Neural Network (GRNN), are trained on a dataset of 121 samples generated using varying patch sizes. The FFBPN model is better with a mean squared error of 3.9506 × 10-5 at epoch 21. The optimized antenna has resonances at 2.448 GHz and 5.8 GHz with S11 < -10 dB, bandwidths of 610 MHz (2.21-2.82 GHz) and 400 MHz (5.61-6.01 GHz), and peak gains of 2.262 dBi and 3.861 dBi, respectively. Measurements on a prototype are in agreement with simulations to confirm the appropriateness of the design for wireless medical devices, such as wearable sensors and telemedicine systems.