Abstract:

A linear stationary control object, the model which is defined by a linear differential or difference equation with exact up to unknown parameters. It is considered that on the basis of experiment data "input -output" are obtained. The task is consisting to identification the parameters of the given equation, providing sufficient proximity of the output variable of the controlled object and its model. For the solve the formulated task is entered penalty function, defined the norm of errors identification. Be offered a new criteria ratio guaranteeing achievement of the purpose of identification. On the basis given ratio of initial task to the brought of minimizing penalty function in the parameter space of the object model. Thus the circuit of the set-up model is used, and also receive the differential equation of self-adjustment (adjustment) parameters. As an assessment of required parameters used the set solutions of the equations of adaptation.

Keywords: control object, object model, parametrical identification, criteria ratio, algorithm of identification