Study of nonparametric models of dynamic systems

The problem of identification and construction of models of stationary linear dynamic objects with delay in conditions of non-parametric uncertainty is observed. The goal of this work was to study models of dynamic objects of the high order. The problem of construction of a model of an object in a state characterized by non-zero initial conditions is considered. Duhamel and Cauchy-Lagrange integrals are used as models of dynamic objects. Non-parametric restoration methods of regression curves are used for construction of models of characteristics of objects. The results obtained demonstrate the possibility of construction of adequate quality models of stationary linear dynamic objects with delay in conditions of non-parametric uncertainty. The results can be applied in the field of automatic process control of production or processing and other fields, where tools for automatic control are used.