Feng Feng, Tianjin University (Special session 13)

Invited Talk: Feng Feng, Tianjin University

Special session 13:  Modelling, Optimization and Applications of Antennas and Microwave Devices


Short Bio: 
Feng Feng received the B.Eng. degree in Tianjin University, Tianjin, China, in 2012, and the Ph.D. degree in the School of Microelectronics at Tianjin University, Tianjin, China, and the Department of Electronics at Carleton University, Ottawa, ON, Canada, in 2017. From 2017 to 2020, he was a Postdoctoral Fellow in the Department of Electronics at Carleton University, Ottawa, ON, Canada. In 2020, he joined the School of Microelectronics at Tianjin University, Tianjin, China, where he is currently an Associate Professor.

He is the Senior Member of the Chinese Institute of Electronics. He serves as a committee member of IEEE MTT-2 technical committee – Design Automation Committee. He also serves as a reviewer for many scientific publications, including IEEE Transactions on Microwave Theory and Techniques, IEEE Microwave and Wireless Components Letters, IET Microwaves, Antennas & Propagation Journal, International Journal of RF and Microwave Computer-Aided Engineering, and International Journal of Numerical Modelling: Electronic Networks, Devices and Fields.

Title: Recent Advances in ANN for Fast Parameterized Modeling and Optimization

Artificial neural networks (ANN) are information processing systems with their design inspired by the studies of the ability of the human brain to learn from observations, and to generalize by abstraction. Researchers have investigated a variety of important applications utilizing the ability of ANN to perform modeling and optimization of microwave components and circuits. ANN has been a recognized vehicle for the electromagnetic (EM) parameterized modeling, i.e., modeling EM behaviors with geometrical parameters as variables. EM parameterized modeling is important for fast EM design optimization. Direct approach to EM design optimization is usually computationally expensive because it requires repetitive EM simulations due to adjustments of the values of geometrical parameters. ANN becomes an efficient method for EM parameterized modeling by learning the relationship between EM responses and geometrical parameters.
An advanced ANN parameterized modeling approach, which combines neural networks and transfer functions (neuro-transfer function or neuro-TF), has been developed to perform parameterized modeling of EM responses. The neuro-TF method is in general an efficient knowledge-based method which uses transfer functions as the prior knowledge when suitable equivalent circuit models/empirical models are not available. The trained neuro-TF models provide fast answers of EM behaviors of microwave components when geometrical parameters are repetitively changed and can be used in high-level design.