Exploiting Meta-Information to Train Sensor Operators (EMITS)

A sensor operator supporting Air Force intelligence operations is often inadequately prepared to reason about qualifiers of incoming information(meta-information) resulting in poor situational awareness and decision-making. To improve a sensor operator’s ability to support the intelligence process, we propose to design, develop, and evaluate a system for Exploiting Meta-Information to Train Sensor Operators (EMITS). Under our Phase I effort, we designed prototype intelligent agent-based training components in a representative operational context. We performed an analysis of sensor operator reasoning about meta-information, and identified how experts perceive, understand, and act upon meta-information. This analysis drove the design of our demonstration software, which intelligently presents training materials to address the role of meta-information in the intelligence process. Under our Phase II effort, we propose to develop the EMITS system by: (1) extending our analysis to support the development of parameterized training vignettes; integrating (2) intelligent methods for assessment of trainee performance and presentation of pedagogic materials with (3) DoD simulation environments; (4) evolving the software system from a proof-of-concept prototype into a complete training product ready for integration into DoD training systems; (5) validating our methodology and refining the training system; and (6) identifying and pursuing Phase III transition and commercialization opportunities.