This paper presents an overview of Neuro-space mapping (Neuro-SM) approach and its application to nonlinear device modeling. The Neuro-SM approach addresses the situation where an existing device model cannot fit new device data well. By modifying the current and voltage relationships in the model, the Neuro-SM produces a new model exceeding the accuracy limit. This paper describes several Neuro-SM techniques incorporating static Neuro-SM, advanced static Neuro-SM, and dynamic Neuro-SM techniques for microwave device modeling. A real 2 × 50 gatewidths GaAs pseudomorphic high-electron mobility transistor (pHEMT) modeling example is used to illustrate the accuracy and efficiency of dynamic Neuro-SM.

Additional Metadata
Keywords Neuro-SM, nonlinear device modeling, Optimization methods
Persistent URL dx.doi.org/10.1109/LAMC.2016.7851276
Conference 1st IEEE MTT-S Latin America Microwave Conference, LAMC 2016
Citation
Liu, W. (Wenyuan), Zhu, L. (Lin), Na, W. (Weicong), & Zhang, Q.J. (2017). An overview of Neuro-space mapping techniques for microwave device modeling. In LAMC 2016 - IEEE MTT-S Latin America Microwave Conference. doi:10.1109/LAMC.2016.7851276