The potential disinteroperabilities and inefficiencies of current fusion architectures can be mitigated in a number of ways including reduced, coordinated, and standardized coupling of components. This proposal is for an ontology-based fusion architecture that will enable fusion agents to operate in independent, yet coordinated, manners. It is based on recent research in the area of ontology-based fusion in the international fusion community. Yet we will show how the approach is also well-grounded in actual DoD fusion systems and algorithms experience, indeed the idea of an ontology-based fusion resulted from many years of fusion project experience. What is planned for the Phase 1 effort is two parallel tracks, (1) develop the theory of ontology networks as a basis for inference; and (2) proof-of-concept demonstrations of the efficacy of the approach. For the proof-of-concept demonstrations, we plan scaled down experiment to verify the feasibility of independent fusion modules (inferlets) to operate within an spreading activation network based on an ontology. We will use a subset of the C2 Core ontology to support experimentation with already-developed algorithms for, (1) tracking filters, (2) association and assignment, and, if the Phase 1 option is exercised, (3) ESM/ELINT classification and identification.
Keywords: Fusion, Ontology, Inference, Open Architecture