Non-parametric models of stochastic systems with delay

Abstract

The paper considers an identiЇcation problem for stochastic static systems with delay. The identiЇcation problem is investigated in a broad sense, specially under conditions of non-parametric uncertainty, i. e. in the case when parametric structure of the model is known to within parameters vector. The new classes of non-parametric estimations of regression line of observations is proposed. The paper gives the corresponding theorems of convergence for non-parametric esti- mation and models.