CHARLES RIVER ANALYTICS, INC. — Department of Defense SBIR Phase I: N181-018

Description

Mission planning is critical to successful deployment of military power. Mission planning is a complex task that requires the detailed orchestration of many subcomponents and elements of human performance in varied, real world environments. This task is becoming yet more complex and challenging as the DoD begins to explore missions that include both manned vehicles and unmanned platforms and incorporate multi-domain air, ground, and sea elements. Mission planning tools and processes do not yet scale to these future challenges, and mission planners may soon be overwhelmed by this complexity. To meet this need, we propose to design and demonstrate Learning Upgrades to Mission planning ENvironments (LUMEN). LUMEN features: (1) a deep learning suite applying deep learning approaches to mission planning problems and mission planner support to improve the efficiency and effectiveness of mission planning workflows; (2) automatically generated causal DNN explanations to enable appropriate trust and understanding of deep learning recommendations; (3) a narrative-based human machine interface (HMI) design to enable mission planners to effectively use these models; and (4) a joint mission planning system (JMPS) integration and security classification layer to enable interoperability with current and future mission planning systems.