Determination of Channel Parameters of FinFET using Artificial Neural Networks in Current Mirror Applications

Short channel effects are seen in planar MOSFET and CMOS transistors with a size smaller than 50nm. The Fin Field-Effect Transistor (FinFET) was developed in response to MOSFET transistors as the size decreased below 50nm. This has resulted in reduced power consumption and less short channel effects. The transition from MOSFET transistor to FinFET technology brings challenges such as complex design and rapid increase in resistance. This challenge will require the designer to determine the values of the FinFET transistor channel parameters entirely by his/her own experience. The aim of this study is to design an artificial neural network model to obtain acceptable and successful channel width and channel length values of the FinFET transistor to be determined by the circuit designer. For the test of the designed model, cascode current mirror and Wilson current mirror are considered in this study.