Scientists and software engineers at Charles River Analytics created an intelligent tutoring system as a key part of an overarching VMT solution developed by DiSTI Corporation, the prime on the effort.

“We developed intelligent tutor and procedure monitoring software for the student and instructor workstations,” said Sean Guarino, Principal Scientist at Charles River. “The software provides an advanced error tracking capability that dynamically classifies and reclassifies mistakes made by the trainees while rating their performance, providing guidance for improvement.”

To develop the system, the team built on previous work in small-business innovative research (SBIR) efforts, including a skill-tree framework that drives adaptive training in LASTS, and performance tracking and analysis developed as part of an adaptive, game-based environment in MAGPIE.

Charles River modeled 66 essential UH-72 maintenance tasks in detail, capturing the individual steps that must be performed in each scenario, as well as meta-information detailing how to dismantle, repair, and reconstruct the equipment. A robust error tracking framework analyzes user performance, determines when errors occurred, and reclassifies the errors based on the actions of the trainee to provide an automated understanding of the specific mistakes made by the trainee (e.g., differentiating missed, added, incorrect, and out-of-order steps). Summary data analysis tools support feedback within the broader VMT training system, reporting on trainee performance and errors to maximize skill improvement.