Article

Article title SUFFICIENT CONDITIONS FOR REDUSING-ORDER OF LINEAR FILTRATION FOR JUMP-WISE PROCESSES
Authors E.Ya. Rubinovich
Section SECTION V. CONTROL SYSTEMS
Month, Year 03, 2014 @en
Index UDC 519.283
DOI
Abstract It is well known that an optimal (in the mean square sense) linear filter for jump-wise Markov process with N-dimensional state-vector and known transition matrix has a dimension N-1. In cases when a transition matrix not depends on time it is possible sometimes to reduce the filter dimension. The paper deals with sufficient conditions for redusing-order of linear filtration for jump-wise Markov processes with N states. An example of possible reduction from N-1 up to dimension one is given. In applications the filters under consideration are used for state estimation in tracking problems with maneuvering targets where a maneuver is described by jump of some jump-wise Markov process with N-dimensional state-vector and known transition matrix.

Download PDF

Keywords Markov processes; filtering; dimension redusing.
References 1. Рубинович Е.Я. Робастная линейная фильтрация скачкообразных процессов // Известия ЮФУ. Технические науки. – 2013. – № 3 (140). – С. 149-155.
2. Krichagina N.V., Liptser R.Sh. and Rubinovich E.Ya. Kalman filter for Markov processes // In: Statistics and control of stochastic processes. – New York: Publ. Div., 1985. – P. 197-213.

Comments are closed.