Temporal alignment in soft information processing

In a counterinsurgency environment, patrol or spot reports and human observations stream into a soft information fusion node in arbitrary temporal order. This paper discusses the use of discrete temporal features and interval logic for temporal alignment in soft information. This aligned sequence can be presented to a graph matching method for situational understanding. Temporal transformations will provide a “present” tense time for incoming messages, which will be represented in a graphical form. The use of the OWL-Time ontology, along with a rule based inference engine JENA, enables messages to be placed in linear order. Given such information, it is feasible to match given scenarios of interest which contain temporal specific requirements. This is accomplished using a temporal version of graph matching. Extensions on this research will include considerations of uncertainty in time through fuzzy representations.