Terahertz Antenna Design Using Machine Learning Assisted Global Optimization Techniques

In the terahertz (THz) frequency band, high path loss necessitates designing the antenna with high-gain characteristics, making it challenging to obtain an optimal design solution. This paper investigates the feasibility of artificial intelligence (AI)-driven antenna design techniques to address the challenges, specifically the Surrogate Model- assisted Differential Evolution for Antenna Synthesis (SADEA-I) algorithm. SADEA-I has been employed for the first time (to the best of our knowledge) to optimize THz antennas. The simulation results show that SADEA-I is more effective than conventional design methodologies. Additionally, it takes only 30 hours to perform the global optimization, and it is a fully automated process and does not require any initial design.