Title: Using generative model for intelligent design of dielectricresonator antennas
Abstract: In the advancing field of 5G technologies, particularly at the 60 GHz band, dielectricresonator antennas (DRAs) stand out for their low conduction loss and high radiation efficiency. However, the traditional design process for DRAs, predominantly reliant on intuitive reasoning and trial‐and‐error methods, is notably inefficient and resource‐intensive. Addressing this critical challenge, our research introduces a pioneering approach: a generative adversarial network (GAN)‐based model specifically tailored for automating D
RA structure design. This novel model represents the first of its kind in the domain, marking a significant departure from conventional methods. Our GAN model uniquely integrates a simulator for DRA modeling and a generator for DRA structure design, streamlining the design process. To effectively train this model, we created a simulated data set comprising pattern–annotation pairs of geometric shapes andS11parameters. This data set enabled the GAN to capture the intrinsic principles underlying DRA design. The practical impact of our model is profound; it significantly expedites the DRA design process, aligning it more closely with specific user requirements while conserving valuable time and resources. This breakthrough approach not only enhances the efficiency of DRA design but also sets a new standard in antenna technology development for future wireless communications.
Bio: Mingdian Liu is a Ph.D. candidate in Electrical Engineering at Iowa State University. He is also a Master student in Computer Science and minor in Statistics. He is working as a student member in Dependable Data Driven Discovery (D4) Institute and Center for Nondestructive Evaluation (CNDE) at ISU. His Ph.D. research focus on the application of AI for Science and Engineering, such as computational electromagnetics and CT reconstruction under the tutelage of Prof. Meng Lu and Prof. Jiming Song. He has internship experiences at Snap Research, Amazon, and OPPO US Research Center where he developed deep learning models for applications in Generative Models, Multimodal Learning, and Pose Estimation. He is looking forward to any research collaborations for the application of deep learning in computational electromagnetics and antenna design.
Please click this URL to start or join. https://iastate.zoom.us/j/96810972944?pwd=SVVLWlY2cVdZYXhxWWg4ZHF1cVdSZz09
Or, go to https://iastate.zoom.us/join and enter meeting ID: 968 1097 2944 and password: 334840
Join from dial-in phone line:
Dial: +1 309 205 3325 or +1 312 626 6799
Meeting ID: 968 1097 2944
Participant ID: Shown after joining the meeting
International numbers available: https://iastate.zoom.us/u/aqUgrVklM