A Novel Dynamic Neuro-Space Mapping Approach for Nonlinear Microwave Device Modeling

作者：Zhu, Lin^*; Zhang, Qijun^*; Liu, Kaihua^*; Ma, Yongtao^*; Peng, Bo^*; Yan, Shuxia^*

来源：IEEE Microwave and Wireless Components Letters, 2016, 26(2): 131-133.

DOI：10.1109/LMWC.2016.2516761

• Dynamic neural networks (DNN)
• neuro-SM
• nonlinear device modeling
• optimization methods

This letter presents a novel dynamic Neuro-space mapping (Neuro-SM) technique for nonlinear device modeling. This is an advance over the existing static Neuro-SM which aims to map a given approximate device model towards an accurate model. The proposed technique retains the ability of static NeuroSM in modifying the effects of nonlinear resistors and current sources. The proposed technique can also make up for any capacitive effects and non-quasi-static effects that maybe missing in the given model, which is not achievable by the existing static NeuroSM. In this way, the dynamic Neuro-SM model can exceed the accuracy limit of the static Neuro-SM. The validity and efficiency of the proposed approach are verified through two high-electron mobility transistor (HEMT) modeling examples.

• 出版日期2016-2
• 单位天津大学