Identification of models for discrete linear systems with variable, slowly varying parameters

The article suggests a solution to the problem of identification of models for discrete linear systems with variable, slowly varying parameters. It is established that the solution can be obtained via minimization of empirical loss functions with equality restrictions defined on a sequence of local intervals. It is shown that the identification is provided with the help of a sequence of functional polynomials satisfying the required conditions of smooth joining. In particular, periodically varying parameters of discrete systems associated with a digital resonator are identified with the use of polynomials of exponential functions.