Identification of Analytical Device Models (Identificatie van Analytische Modellen van Elektronische Componenten).

摘要

Analytical device models play an important role in the design and analysis of electronic circuits. Analytical device models are the natural choice for applications such as process characterization and the design of integrated circuits and devices. The accuracy with which a device model can predict the actual behavior of a device depends on the proper definition of the model parameters. In this thesis we present a unified approach to the identification of analytical device models. The identification methods considered are all based on data-fitting techniques; the model parameters are determined by optimizing an identification criterion that expresses the difference between the observed device behavior and the device behavior that is predicted by the model. The choice of the identification criterion is based on assumptions about the characteristics of the model and the observation errors. Hence, the acceptability of an identification method depends on how realistic these assumptions are when applied to practical identification problems.