Reconnaissance Target Detection

In a previous paper, we have described a rule-based approach to the aggregation and enrichment of incomplete situational pictures that represent so-called 'perceived truths' [1]. The approach enables the derivation of all complete situational pictures that are compatible with a particular perceived truth. These alternative complete pictures can be seen as conflicting hypotheses on the ground truth. The aim of further reconnaissance is to eliminate such hypotheses and thereby narrow down the space of possible situational pictures. To this end, an information-theoretical concept of 'value of information' (VoI) can be operationalized: information is more valuable, when it provides a better contribution in eliminating false hypotheses. Reconnaissance targets have to be selected according to the value of the information they can provide. In practice, hypotheses are clustered regarding their predictions on the location of specific Battle Space Objects (BSOs). Conflicts between the clusters are translated into decision questions: are BSOs of type X located in area Y? Reconnaissance targets are selected to answer such questions. In this paper, we will elaborate on the concept of reconnaissance as hypothesis testing and introduce a first demonstrator for the automated determination of reconnaissance targets based on enriched situational pictures. This paper was originally presented at the NATO Science and Technology Organization Symposium (ICMCIS) organized by the Information Systems Technology (IST) Panel, IST205-RSY - the ICMCIS, held in Koblenz, Germany, 23-24 April 2024.