Rapid Development of Scenario-Based Simulations and Tutoring Systems

Abstract

§Low earth orbit constellations significantly increase the burden for ground-based systems management. Efficient, accurate management requires that ground station and satellite operators are well-trained. Scenario-based training, in which trainees practice handling specific situations using faithful simulations of the equipment they will use on the job has proven to be an extremely effective method for both training and certification. Effectiveness is increased when it is coupled with intelligent tutoring. Intelligent tutoring systems (ITS) improve on training simulations by assessing the trainee’s actions and providing individualized guidance. Development of training simulators often lags behind development of the operational systems, with the result that simulators are often not available or up to date when they are most needed (e.g., when training operators on new or modified systems). Also, practical deployment of intelligent tutoring systems is hindered by the time and expense required to prepare the needed knowledge representations. We report on the development of tools to address both these issues. Three strategies enable rapid development. For both simulation and tutoring, the first strategy is to develop systems that are scenario-specific. This strategy permits the systems to be simpler and easier to develop. The second strategy is to employ easily manipulated representations for the simulation behavior and tutoring knowledge, and provide graphical user interfaces to edit these representations that facilitate development by training instructors with little outside support. The third strategy is to minimize the need for development. For simulation, this means maintaining a high degree of fidelity with little programming by incorporating screen captures of the operational system, and by enabling integration of existing simulation codes. For tutoring, this means capturing the bulk of the tutoring knowledge by simply recording the actions of instructors as they demonstrate the correct manner of handling the scenario. These strategies are embodied in two systems developed by Stottler Henke for the Air Force. TaskSim is a simulation framework to facilitate rapid development of scenario-based training simulations. Task Tutor Toolkit™ is an intelligent tutoring system that enables instructors to create scenario-specific tutors through a three-step process: demonstrate, generalize and annotate.