This paper presents a general method combining microwave empirical/equivalent model with artificial neural networks for efficient modeling of microwave components. The method, called Generalized Knowledge-Based Neural Network (GKBNN), unifies several existing methods and provides increased model accuracy and extrapolation capability, even if the training data is limited. It is applied to microwave device and transmission line modeling for high frequency/high speed circuit design.

Additional Metadata
Conference 2002 International Symposium on Antenna Technology and Applied Electromagnetics, Antem 2002
Citation
Zhang, L. (Lei), Ding, R.T., & Zhang, Q.J. (2002). Generalized knowledge-based neural network for microwave modeling. In Antem 2002 - International Symposium on Antenna Technology and Applied Electromagnetics, Conference Proceedings.