Transistor Model Building for a Microwave Power Heterojunction Bipolar Transistor

This article describes the detailed development of a large-signal transistor model for a microwave power heterojunction bipolar transistor (HBT). Based on the neurospace-mapping (Neuro-SM) technique, this model uses neural networks to map a coarse model space represented by the existing model simulations onto the fine model space represented by device measurements. This article tied for first place in the Microwave Transistor Modeling Student Design Competition at the 2014 IEEE International Microwave Symposium (IMS2014) held in Tampa, Florida.