Development of time and workload methodologies for Micro Saint models of visual display and control systems

Abstract

The Navy, through its Total Quality Leadership (TQL) program, has emphasized the need for objective criteria in making design decisions. There are numerous tools available to aid human factors engineers meet the Navy’s need. For example, simulation modeling provides objective design decisions without incurring the high costs associated with prototype building and testing. Unfortunately, simulation modeling of human— machine systems is limited by the lack of task completion time and variance data for various objectives. Moreover, no study has explored the use of a simulation model with a Predetermined Time System (PTS) as a valid method for making design decisions for display interactive consoles.

This dissertation concerns the development and validation of a methodology to incorporate a PTS known as Modapts into a simulation modeling tool known as Micro Saint. The operator task context for the model was an interactive displays and controls console known as the AN/SLQ-32(V). In addition, the dissertation examined the incorporation of a cognitive workload metric known as the Subjective Workload Assessment Technique (SWAT) into the Micro Saint model.

The dissertation was conducted in three phases. In the first phase, a task analysis was performed to identify operator task and hardware interface redesign options. In the second phase data were collected from two groups of six participants who performed an operationally realistic task on 24 different configurations of a Macintosh AN/SLQ-32(V) simulator. Configurations of the simulated AN/SLQ-32(V) were defined by combinations of two display formats, two color conditions, and two emitter symbol sets, presented under three emitter density conditions. Data from Group 1 were used to assign standard deviations, probability distributions and Modapts times to a Micro Saint model of the task. The third phase of the study consisted of (1) verifying the model-generated performance scores and workload scores by comparison against scores obtained from Group 1 using regression analyses, and (2) validation of the model by comparison against Group 2.

The results indicate that the Modapts/Micro Saint methodology was a valid way to predict performance scores obtained from the 24 simulated AN/SLQ-32(V) prototypes (R² = 0.78). The workload metric used in the task network model accounted for 76 percent of the variance in Group 2 mean workload scores, but the slope of the regression was different from unity (p = 0.05). The statistical finding suggests that the model does not provide an exact prediction of workload scores. Further regression analysis of Group 1 and Group 2 workload scores indicates that the two groups were not homogenous with respect to workload ratings.