The Department of Defense is developing plans to transform training and education; providing more dynamic, capabilities-based training programs in support of national security requirements associated with today's strategic environment. Capabilities-based training must facilitate the warfighter's ability to respond quickly and adapt effectively when faced with rapidly evolving, asymmetric warfare. Training tools such as future threat generation systems, modeling techniques, and performance measurement methods themselves must be flexible enough to cultivate responsiveness and adaptability in current and future warfighters.

The development and execution of adaptive threat generation systems and rapid modeling techniques within applied research and training environments poses many methodological and integration challenges. The Air Force Research Laboratory, Warfighter Readiness Research Division (711 HPW/RHA) is assessing the critical issues facing present day threat generation systems, threat models, and the extent to which current modeling frameworks/architectures can provide military training with accurate/credible models of human behavior. The objective is to explore the efficacy of incorporating real-time performance data into architecture(s) to develop models (e.g., realistic constructive adversaries) for use in a threat generation/adaptive training systems. Also, the possible types of assessment criteria (e.g., verification and validation) needed to define levels of successful integration.

This paper will briefly examine the critical issues facing future warfighters, present a high-level summation of the current research and practices associated with threat generation/adaptive systems and their current use of modeling techniques, and will discuss salient lessons learned. Finally, the team will suggest directions on how these tools can be used to better prepare the warfighter to answer the dynamic challenges of future warfare. This research can serve as a foundation in the development of more adaptive, capabilities-based training tools (e.g., less-scripted, more dynamic training scenarios with realistic constructive forces) for use in training and ops communities such as Live Virtual Constructive (LVC) environments.