Analysis of Shiphandlers' Eye-Gaze and Simulation Data for Improvements in COVE-ITS System

An intelligent tutoring system (ITS) is paired with U.S. Navy immersive virtual trainer for shiphandling called Conning Officer Virtual Environment (COVE) to monitor students performance and provide spoken feedback. The feedback can be improved by incorporating the cognitive state of shiphandlers through analysis of their attention-allocation patterns from eye-tracking data. This thesis research contains a pilot study directed toward analyzing the eye-tracking data of expert and novice shiphandlers. We examined the relationship between a shiphandlers experience level, attention-allocation patterns, and performance during a simulated shiphandling exercise. Five novice and four expert shiphandlers from the U.S. Navy Surface Warfare Officers School participated in the study. Our analyses indicates expertise differences in shiphandling performance, general eye-tracking measures, scan transitions, as well as time distribution between different areas of interest. The experts superior shiphandling performance was linked with having targeted and tight attention-allocation patterns that focused only on the relevant areas of interest. Novices attention-allocation patterns were highly scattered and irregular. Results suggest that incorporating the ideal attention-allocation patterns of the experts into the ITS could improve its feedback to novice shiphandlers by telling novices where they should look and when. The study is based on a small sample size; therefore, further data collection should be performed to confirm the results.