Explicit parameter identification for a class of nonlinear input/output differential operator models

The system parameter identification problem for a class of nonlinear input-output differential operator models is formulated in terms of an explicitly defined least squares cost function. Central to the formulation is a discrete-frequency format brought about by the introduction of a set of explicitly defined Fourier-based modulating functions within the framework of the classical Shinbrot method of moment functionals. The underlying calculations needed to set up the least squares problem differ according to whether the differential operator model is of the exact or inexact category. Examples are presented of each such category.<>