Flexible and Live Adaptive Training Tools (FLATT)

Intelligent Tutoring Systems (ITS) and adaptive training solutions promise the ability to provide adaptive training focused on the specific needs of the trainee and their performance. To achieve adaptive training, software changes are often required to adjust the adaptions or the criteria under which they are applied, which is a lengthy process on the order of weeks or months. With advanced technologies in human-wearable sensors, data can be used to develop new models of learners, and determine effective training adaptions that are appropriate for each learner state. Trainers can use this data to increase the effectiveness of training solutions, by characterizing a learner’s state and then adapting the training. Unfortunately, there is no simple, generalized mechanism available for trainers to adapt an on-going training exercise based on the real-time learner state. Researchers and trainers require an easy-to-use toolset to capture and visualize information on a learner’s state, and then act on that data to adapt the training in real time. The Flexible and Live Adaptive Training Tools (FLATT) SBIR Phase I research explored the state of existing ITSs, virtual training environments (VTEs), physiological sensors, trainee state customizations (TSCs), visual rule authoring user interfaces and rules engines in order to inform the FLATT system design. FLATT will provide researchers and instructors a flexible tool to easily adapt virtual training exercises in real-time based on the trainee’s state and actions.   During Phase II, Dignitas will develop the FLATT prototype based on our Phase I system design as a user friendly, rules-based authoring tool that leverages different data sources and applies various TSCs. Based on the outcome of our Phase I analysis, we selected the Generalized Intelligent Framework for Tutoring (GIFT) as the ITS at the core of FLATT. Our prototype will integrate with VBS3, and we will target our initial transition efforts towards Games for Training (GFT) and Army school houses. As STE is defined and matures, we will monitor and work to transition FLATT into STE virtual training systems in order to benefit future objectives.  FLATT will be incorporated into the GIFT product as well, which is released annually via an open source license, providing benefit to over 2,000 GIFT users. Research conducted in Phase I established a strong foundation for rapidly building a prototype that integrates with various VTEs. The FLATT prototype developed in Phase II will provide many benefits, including a flexible user experience in a tool with a Service Oriented Architecture (SOA) that enables easy integration with different technologies or solutions. Researchers will be able to develop new models of learners and determine effective training adaptations that are appropriate for each learner state. Trainers will be able to increase the effectiveness of training solutions by characterizing a learner’s state and then adapting the training.