Damage Control Skills Automated Assessment

Assessment, Methodology, Technology

In a multi-year research project funded by the Office of Naval Research, CRESST developed a software architecture for automatically assessing cognitively complex tasks within games and simulations. The focus of this research was the development of a conceptual framework for automated performance assessment that can be generalized to a variety of domains. The goal was to provide a generic model of assessment that can be incorporated into both new and pre-existing computer-based simulations that depict cognitively complex scenarios.

As part of this effort, CRESST developed a 3-D, computer-based simulation depicting the interior of a Navy ship, in which the player must assume command of a damage control repair locker. In this role, the player is tasked with identifying, responding to, and managing the containment of a variety of fire and flood casualties that break out in various areas of the ship. Using Bayesian inferencing, the game’s automated assessment engine analyzes all meaningful player actions and decisions, and provides estimates of the player’s proficiency in several key areas including: personnel safety, resource management, compartment integrity, communications, situational awareness, and decision making.

Project Director(s):

Alan Koenig, John Lee