A Model of the Relative Effects of Key Task and Display Design Parameters on Training Task Performance

 

Charles J. Lloyd
L-3 Communications, Link Simulation & Training

Logan Williams,  Byron Pierce
711 HPW/RHA, Air Force Research Laboratory

From the Proceedings of the 2011 IMAGE Society Annual Conference, Scottsdale, Arizona.

A Model of the Relative Effects of Key Task and Display Design Parameters on Training Task Performance A Model of the Relative Effects of Key Task and Display Design Parameters on Training Task Performance

Abstract

This paper describes the design and initial validation of a model of the relative effects of key task and display design parameters on training task performance. The display system design parameters include pixel hold time, pixel pitch, luminance, contrast, and noise. The task and observer parameters include target size and range, angular velocity (of the image), target contrast, and observer capability (e.g., acuity).

This model was developed as part of the Immersive Display Evaluation and Assessment Study (IDEAS) program for the USAF. The model was designed to be used by display system acquisition professionals who must develop defensible display system requirements that are based on the ability of the display system to support the training planned for the system. The model was also designed for those display systems professionals in the supplier community who wish to steer their product designs in the direction of maximum utility to the USAF.

The initial validation results indicate the model accurately summarizes the findings from several published evaluations that employed tasks generally representative of flight simulation training.  The model was then used to make specific predictions of the effects of 220 combinations of four parameters (hold time, velocity, pitch, and luminance) on the range at which pilots could reliably identify fighter sized aircraft.  A formal evaluation of the effects of these parameters was conducted and the results of this evaluation are compared with the predictions.  A high correlation between model predictions and the results of this evaluation was obtained. The details of this evaluation are presented in a companion paper published at this conference.