On nonparametric models of linear dynamic objects

The problem of identification of stationary linear dynamic objects in a non-parametric uncertainty is considered. One of the classical methods of nonparametric identification of the characteristics of objects is studied. A generalization of this method allowing the identification of the characteristics of objects, with smooth, feasible in practice signals is proposed. The solution to the problem of low noise-immunity of the method is proposed. The method is illustrated by the numerical experiment. The applicability of the method of objects being identified by means of passive experiments is considered. The models based on the results of identification experiments that extend the class of models of stationary linear dynamic objects are proposed.

Keywords

Authors

References