On nonlinear filtering problems for discrete time stochastic processes

From MaRDI portal
Publication:2583057





After Kalman-Bucy's work for the linear filtering problem for Gaussian systems of Markovian type, the extended Kalman filter has been derived based upon Taylor approximations of nonlinear systems of Markovian type. However, it is said that the linearization of the nonlinear system by Taylor approximations provides an insufficiently accurate representation in many cases. The purpose of this paper is to develop linear causal analysis for systems consisting of two flows in a real inner product space. As application, an algorithm for calculating the nonlinear filter for a discrete stochastic system consisting of a signal process and an observation process without Dobrushin-Minlos' regularity condition and based on the theory of \(KM_2O\)-Langevin equations is obtained.




Cited in
(27)








This page was built for publication: On nonlinear filtering problems for discrete time stochastic processes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2583057)