Konferenční příspěvek

Straka O., Šimandl M.

Proceedings of the Eighth IASTED International Conference on Intelligent Systems and Control

ACTA Press

Anaheim, USA

2005

96 - 101

0-88986-517-5

Ing. Ondřej Straka, Ph.D. (FAV Západočeská univerzita, katedra kybernetiky)

stochastic systems, nonlinear estimation, Gaussian sum

clanek_IASTED05_OS.pdf [file]

prezentace_IASTED05_OS.pdf [file]

State estimation of the non-Gaussian systems by the Gaussian sum method is treated. The distance-based pruning technique is designed for an approximation of the filtering probability density function given by a weighted sum of Gaussian distributions. The technique measures significance of each term of the sum using the Lissack-Fu distance between the approximate filtering probability density function and the filtering probability density function and prunes the insignificant terms. The paper also proposes a thrifty implementation of the developed technique. The distance-based pruning technique provides high approximation quality in comparison with other approximation techniques, moreover it achieves low computational demands as it is illustrated in a numerical example.