MAGPIE is our game-based training framework for developing intelligent virtual training systems.
MAGPIE combines an integrated suite of efficient content authoring tools, models of trainee skill and motivation, and a game adaptation engine to dynamically deliver game-based maintenance training that is responsive to individual learning needs, performance, and instructor guidance.
MAGPIE includes an intelligent tutoring system (ITS) that provides both real-time performance assessment and an ability to construct proficiency models to understand the trainee’s current skills and skill gaps. MAGPIE’s ITS adapts training material, student help, and instructor support to the needs of the individual user. We deployed MAGPIE at Sheppard Air Force Base, where instructors and students are actively using it for F-15E avionics maintenance training.
Maintenance Training based on an Adaptive Game-Based Environment using a Pedagogic Interpretation Engine (MAGPIE)
The US Air Force instructs aircraft maintainers through a combination of live exercises and classroom training. These approaches can be costly and time consuming, and are limited in their ability to demonstrate a comprehensive range of realistic maintenance situations or to provide compelling, tailored training to individual students. In contrast, game-based training can provide a powerful, personalized, and motivating approach for training essential aircraft maintenance skills. However, updating and adapting existing maintenance training content to support game-based delivery methods is an expensive undertaking. To address this challenge, the Air Force is seeking more efficient and cost-effective tools for constructing game-based maintenance training.
The Charles River Analytics Solution
Our scientists and software engineers designed and developed Maintenance Training based on an Adaptive Game-Based Environment using a Pedagogic Interpretation Engine, or MAGPIE. MAGPIE combines an integrated suite of efficient content authoring tools, models of trainee skill and motivation, and a game adaptation engine to dynamically deliver game-based maintenance training that is responsive to individual learning needs, performance, and instructor guidance.
“MAGPIE dynamically adapts both the scenario and the mechanics of the game to maximize training effectiveness over time, providing individual aircraft maintainers with customized training for key tasks,” explained Sean Guarino, the principal investigator of the project at Charles River. “This adaptation is based on the trainee’s performance, instructional design theory, and motivational theory; it helps to ensure that the trainee has a beneficial and appropriate training experience when instructor guidance is limited.”
Mr. Guarino added, “MAGPIE promises to reduce training costs in two ways. First, it provides rapid and intuitive authoring that allows course designers to construct new training content with limited software engineering support. Second, it reduces the cost of training exercises by allowing instructors to manage large classes in a virtual maintenance trainer, reducing necessary time and potential costly errors on live vehicles and partial task trainers.”
View a demo of MAGPIE in this video.
The combination of MAGPIE components enables the Air Force to successfully harness the true power and efficiencies of game-based training for maintenance personnel. Its adaptive game-based environment provides a more powerful and personalized training experience than classroom training. It has the potential to lower costs and improve skills for Air Force personnel maintaining air platforms, as well as other assets. MAGPIE can also benefit other agencies that train personnel using computer-controlled immersive scenarios, such as air traffic controllers, emergency responders, and medical response teams.
The project is sponsored by the Air Force Research Laboratory (AFRL) under contract FA8650-14-C-6506.