Request management using contextual information for classification

Abstract

In a multitarget and multisensor environment a faithful and precise operational situation is needed as much as a fast data acquisition and processing in order to make reliable and reactive decisions. From this perspective, we introduced in our last paper the separation degree, which is a measure of discrimination between two fuzzy sets. It is used as a criterion to obtain an order among the target's attributes or among the sensor's modes. Thus it helps the choice of the attribute that is the most characteristic for the targeted object, by selecting the most discriminating fuzzy sets, which would give the less ambiguous result. From this perspective we propose a method to select the sensor's mode, which takes contextual information about the targeted object and sensor's cost into account.