Konferenční příspěvek

Punčochář I., Šimandl M.

Proceedings of the 7th International Ph.D. Workshop: Yound Generation Viewpoint

2006

183-189

80-903834-1-6

Ing. Ivo Punčochář, Ph.D. (FAV Západočeská univerzita, katedra kybernetiky)

active change detection, input signal design, optimization

clanek_IPW06_IP.pdf [file]

prezentace_IPW06_IP.zip [file]

The paper deals with an active change detection problem in a stochastic discretetime linear multiple model framework. An active detector decides on changes in an observed system and also generates an auxiliary input signal that should improve change detection. Design of the optimal active detector is formulated as minimization of an appropriate criterion. The general solution is obtained using Bellman’s principle of optimality. The main contribution of the paper is an analysis of active change detection in case of two scalar models and detection horizon of two steps. It is shown that the auxiliary input signal can improve change detection only if two models differ in certain parameters.