Bayesian approach for robust parameter tracking

Abstract

In this paper we study the problem of tracking of time-varying parameters of a dynamical system. The problems also facing at the finite number of expected parameter changes and finite number of possible measurement model. We consider a stochastic model of parameter development with some form of obsolete information forgetting. It will be shown that it is possible to track rapidly time-varying plant parameters using extension of Bayesian viewpoint (continuous and discrete parameters) with requiring the prior information that can improve tracking for abrupt changes.