|Abstract:||Creating interoperability among heterogeneous systems enhances our military's war-fighting capabilities. Differences in hardware, languages, and data models make interoperability hard to achieve. The Object-Oriented Method for Interoperability (OOMI) resolves modeling differences among systems through construction of a Federation Interoperability Object Model (FIOM) used to capture information and tasks shared among systems. The FIOM is constructed in either a bottom-up or top-down fashion using the OOMI Integrated Development Environment (OOMI IDE) and includes both component system and standard representations of the shared tasks and information.
When constructing a federation of interoperable systems, a correspondence must first be established among shared tasks and information before data modeling differences can be resolved. The OOMI IDE uses both semantic and syntactic correlation methodologies for establishing such correspondences. Syntactic correlation is performed using neural networks. Syntactic data concerning the structure and signature of shared information and tasks is used to create discriminator vectors for objects being compared. Neural Networks are used to compare these discriminator vectors to determine the degree of similarity among objects. A ranking of the scores returned from the neural network comparison is used to assist an interoperability engineer in identifying corresponding objects for which modeling differences can be resolved.|