Adaptive Tactical Learning through Automated Scenarios (ATLAS)

The AEGIS Weapon System (AWS) provides a mission-critical radar and weapons capability for Sailors to eliminate threats to our fleet and assets. Operations involve threat classification and decision-making where errors are costly, such as the recent case of fratricide in which the USS Gettysburg shot down a US F/A 18F over the Red Sea. Evolving operations pose greater challenges as US forces consider adversaries using more complex tactics, equipment, and capabilities, including combined threats posed by missiles, crewed aircraft, and drones. For example, the combined missile and air threat posed by near-peer adversaries create an integrated air and missile defense (IAMD) challenge that is not well addressed in current AEGIS training. Similar threats are posed by less advanced adversaries and terrorists. AEGIS teams need more effective training to learn, build, and maintain IAMD skills, and ensure that those skills address the latest adversary tactics and capabilities.ÝCharles River Analytics is pleased to propose to design and demonstrate a system for Adaptive Tactical Learning through Automated Scenarios (ATLAS), an intelligent adaptive training system that automatically generates individualized, novel scenarios to train through a combination of immersive gamification and in-situ AEGIS training. ATLAS provides robust personalized training through a combination of four components: (1) the IAMD Skill Model, which provides a hierarchical computational representation of targeted skills and domain context using our Methodology for Annotated Skill Trees (MAST); this model includes annotations capturing the decision processes that must be trained, performance assessment metrics, and problem challenge templates to drive scenario generation; (2) the Adaptive Scenario Generator, which assesses individual training needs and automatically generates novel customized scenarios to pose tactical problems for developing IAMD decision-making skills; (3) the Multimodal Training Environment, which provides a combination of immersive training that adapts our gamified IAMD training solution, developed for the Missile Defense Agency's (MDA) and US Air Force (USAF) to provide repetition and practice, with in-situ AEGIS training to enable detailed operational practice, united using our KWYNÆ Training Overlay, an application-independent tool for providing real-time training guidance and automatic, objective performance assessment; and (4) the Adversary Update Engine, which refines adversary capability models using Charles Riverís existing machine learning technology to assess AEGIS training and operational exercise data and evolve ATLAS models of adversary capabilities and tactics. These capabilities provide a robust, adaptive, continuous training solution for building and maintaining IAMD decision-making skills.