| Title |
A Physics-Based Neural Modeling for High Frequency Noise of MOSFETs |
| DOI |
https://doi.org/10.5370/KIEE.2021.70.2.354 |
| Keywords |
nanoscale MOSFETs; artificial neural network; radial basis function; high frequency noise; channel thermal nosie; induced gate noise; correlation noise; shot noise |
| Abstract |
A new neural model is presented to predict the dynamic characteristics of high frequency noise in nanoscale MOSFETs with ultrathin gate oxide. The model is formulated by combining physical theories with the differential and integration of the radial basis function (RBF) from an original artificial neural network (ANN). The high frequency noise model includes the channel thermal noise, the induced gate noise, their correlation noise as well as the shot noise generated by the gate leakage current through the ultrathin gate oxide. By training a Fano factor of the shot noise and spatial distribution of the channel thermal noise, this approach forms a physics-based neural modeling that naturally encodes underlying physical theories as prior information. The model exhibits a good performance for predicting the noise behavior at high frequencies. |