FIATSim: Flexible Immersive Adaptive Training Simulations

US Navy sailors deployed aboard ships around the globe protecting the freedom of the seas face an ever-evolving threat landscape that today is heavily focused on Integrated Air & Missile Defense (IAMD). Current IAMD training leverages traditional training methods such as lectures and shore-based simulation facilities, both reliant on predetermined scenarios. This approach creates significant challenges: repetitive scenarios, which can lead to complacency and limit critical thinking and adaptation skills; inflexible scenario pacing, which does not adapt to individual learning rates; limited access to training while at sea; and most critically, an inability to provide real-time analysis and instruction without direct human instructor oversight – particularly during realistic time-compressed scenarios. We propose the Flexible Immersive Adaptive Training Simulations (FIATSim), an artificial intelligence (AI) and machine learning (ML) (AI/ML) powered training system that uniquely addresses these challenges through three key innovations:  A deployable standalone immersive 3D high fidelity virtual reality (VR)-based, AI-powered system that enables individual and team training both at sea and ashore. An interconnected AI system of Learner, Adaptation, and Content modules that dynamically adjusts instructional approaches, threat presentations, and scenario complexity based on real-time performance analysis. A transparent assessment framework that generates both immediate fleet trainee feedback and detailed performance analytics for trainer and trainee debrief.  FIATSim will deliver an immersive, adaptive, autonomous training environment that equips Navy personnel with the adaptable skills and knowledge needed to counter sophisticated and evolving threats while significantly expanding training opportunities for deployed personnel. This system will continuously improve an operator’s performance by enabling repeated practice for individuals and teams that is tailored to their current strengths and weaknesses. FIATSim will be based on a combination of best practices for adaptive learning environments and assessments. The adaptive learning architecture will leverage prior research on effective personalized learning across a variety of domains and diverse digital mediums (Figure 1; Sottilare et al., 2017) that has shown efficacy for both learning and assessment of skills and knowledge (Kulik & Fletcher, 2016). The components of the FIATSim architecture will leverage Evidence-Centered Design (ECD), which is a highly effective methodology for designing assessments that has previously been applied to simulations (Mislevy, 2013). The proven flexibility of this architecture is critical to FIATSim meeting the needs of the Navy’s IAMD systems and the ever-changing scenarios that Navy personnel must be swiftly and effectively prepared to handle.