Proceedings of American Institute of Aeronautics and Astronautics (AIAA) – Infotech@Aerospace, Atlanta, GA (2010)
Situational awareness (SA), which is the general problem of trying to determine what is happening in a situation given the available information, has many applications, both military and civilian. Effective SA requires models that are rich enough to accurately model the types of objects, events, and relationships between them. They must also be able to model the many types of uncertainty inherent in many domains. Current solutions to building fusion and SA models, such as rule-based systems and Bayesian networks tend to fail to model the probabilistic relationships of the domain or to model the structural richness of the domain. Probabilistic Relational Models (PRMs) are a recently developed representation that addresses this issue by allowing representation of combined structural/relational and probabilistic information. PRMs are an object-oriented representation, modeling the world in terms of objects and relationships between them. PRMs also allow uncertainty over the relational structure to be expressed. Using PRMs will allow broad factors and a wide range of information sources to be taken into account for SA, which will assist in producing faster and more accurate analyses.
The work presented here was supported by DARPA contract W31P4Q-08-C-0413 on Probabilistic and Relational Inferences in Dynamic Environments and Air Force contract FA87050-09-C-0090 on Multi-Intel for Space Situation Awareness.
For More Information
(Please include your name, address, organization, and the paper reference. Requests without this information will not be honored.)