Ontological Models to Support Planning Operations

Automation and autonomy technologies, such as automated planning software, are key elements in realizing the vision for space exploration. A fundamental requirement for success with these technologies is that they operate using valid models or ontologies of the application domains. Making ontological information available to automated systems is difficult because 1) domain experts reason in domain terms, not the formal logic of ontologies; 2) the states and configurations of the specific objects in the domain are both voluminous and dynamic, making manual entry and maintenance prohibitive; and 3) the data required, especially state updates, need to be extracted or imported from other disparate systems. This proposal seeks to investigate, design and test a framework for consistent ontological modeling both within and across domains that can be exploited by automated planners currently being developed by NASA's exploration technology program. Specifically we will investigate a modeling framework that provides 1) an ontological representation of domain information in a standard format that can be used by NASA's developing planning software, 2) an interactive editing environment to allow domain experts to construct and maintain the ontological information; and 3) a general, systematic, and maintainable semantic mapping from external data sets into the user-constructed ontology.