Neuro-space mapping technique for microwave nonlinear circuits modeling

Neural networks are been used for the last years for modeling the nonlinear microwave devices as the Pseudomorphic High Electron Mobility Transistor (pHEMT). In this paper the neuro-space mapping (SMNN)technique is used for modeling the DC characteristic of pHEMT, in this approach the Angelov model is used as the coarse model, which can be adjusted using the fine model based on measurement. Good results are obtained for the drain-current, the trans-conductance and for the output conductance.