AI-Driven Design of Microwave Antennas with Case Studies

Akinsolu, Mobayode O. (2019) AI-Driven Design of Microwave Antennas with Case Studies. In: 2019 Automated RF & Microwave Measurement Society" (ARMMS) Conference, 18-19 Nov 2019, Wyboston Lakes, Wyboston.


Download (1MB) | Preview


With the advent of artificial intelligence (AI), the design of microwave devices such as antennas has been expedited in terms of throughput and time-to-market. This is chiefly because design automation via optimization has replaced the use of time intensive manual design techniques which premise on trial and error without any guarantee of successful outcomes. For the rapid design of antennas via optimization, surrogate model-based optimization (SBO) methods tend to be at the forefront due to their efficiency improvement in terms of computational cost. The surrogate model assisted differential evolution for antenna synthesis (SADEA) algorithm family are a class of state-of-the-art SBO methods. In this paper, the use and advantages of the SADEA algorithm family is demonstrated using two cases of real-world antenna design problems as examples. The antenna design problems are the optimization of: a multi-layered compact multiple-input and the multiple-output (MIMO) antenna array for wireless communications and a microwave imaging antenna for ultra wide band (UWB) body-centric applications. For both examples, the SADEA algorithm family obtained very good design solutions within an affordable time and the quality of the obtained solutions are validated by the close agreement which exits between the simulated and measured results of the fabricated and ready-to-use prototypes of the antennas. In one of the cases (the microwave imaging antenna), the performance of the SADEA algorithm family when compared to 2019 Computer Simulation Technology - Microwave Studio (CST-MWS) optimizers (trust region framework (TRF) and particle swarm optimisation (PSO)) is reported. Results from the comparisons show that the SADEA algorithm family obtains very satisfactory design solutions in all runs using an affordable optimization time in each, whereas the alternative optimizers failed in all runs by not meeting the design requirements and/or generating designs with geometric incongruities.

Item Type: Conference or Workshop Item (Paper)
Keywords: Antenna optimization, artificial intelligence, evolutionary methods, PSADEA, SADEA, SADEA-II, Surrogate model-based optimization.
Divisions: Applied Science, Computing and Engineering
Depositing User: Hayley Dennis
Date Deposited: 12 Apr 2021 13:43
Last Modified: 12 Apr 2021 13:45

Actions (login required)

Edit Item Edit Item