Adaptive Localizer Based on Splitting Trees

Abstract : When testing a black box system that cannot be reset, it may be useful to use a localizer procedure that will ensure that the test sequence goes at some point through a state that can be identified with a characterizing set of input sequences. In this paper, we propose a procedure that will localize when the separating sequences are organized in a splitting tree. Compared to previous localizing sequences based on characterization sets, using the tree structure one can define an adaptive localizer, and the complexity of localizing depends on the height of the tree instead of the number of states.