Situation Assessment Fuselet to Aggregate Retrieved Information (SAFARI)

We propose to develop a Situation Assessment Fuselet to Aggregate Retrieved Information (SAFARI) obtained from the Joint Battlespace Infosphere (JBI) repositories to support the development of mission relevant knowledge via Bayesian Belief Network (BN)technology. A BN can be thought of as a graphical program script representing causal relationships among various battlefield concepts represented as nodes to which observed evidence can be posted. Depending on the battlefield situation, a BN is constructedfrom a library of networks to instantiate SAFARI. An if-then type of speculative analysis on the constructed BN identifies relevant information that needs to be retrieved and posted to nodes for assessing the situation. An inferencing process on the samenetwork performs the aggregation of posted evidence to infer high-level battlefield situations. BN technology offers several advantages, including its graphical approach to situation assessment modeling and handling of uncertainty consistent withprobability semantics. At Charles River Analytics we have been successfully applying BN technology for situation assessment tasks in a number of domains, including Major Theater of War (MTW) scenarios, Theater Missile Defense (TMD) domains, and air combatengagements. We believe that SAFARI will significantly enhance the JBI's capability of transforming data into mission relevant knowledge.Commercial applications of the proposed situation assessment fuselet concept based on belief network technology existfor a wide variety of contexts characterized by high rates of information flow and decision-making under uncertainty. Such application areas include operation centers for complex process control (e.g., nuclear power plants, water resource distribution),and rail and air traffic management centers. Vertical applications for these markets would significantly enhance fault diagnosis, improving effectiveness and safety in a number of life-critical applications.