SCOPE - Pilot Workload Estimation Using Control Response: Theoretical Development and Practical Demonstration

Publication Date

1-1-2020

Document Type

Conference Proceeding

Publication Title

AIAA Scitech 2020 Forum

DOI

10.2514/6.2020-2267

Abstract

Previous work by the author referred to a spare capacity estimator that was under development as the “Bedford Estimator”. This paper formally assigns the name Spare Capacity OPerations Estimator (SCOPE) to the metric. Although initially tailored for compensatory tracking tasks using a display (where the tracked error signal is known), SCOPE’s capabilities have been expanded to accommodate various applications. Previously the author showed how control rate and error rate can be used to estimate workload by employing the General Power Law and psychophysical principles. It is demonstrated that the workload estimator implicates the error rate power spectral density, shaped by pilot compensation, as the physical workload stimulus. Analysis of the Crossover Model output within the bandwidth of operation shows that nulling error rate should play a key role in pilot control strategy, hence error rate should be a dominant influence on workload. A model is proposed where the pilot timeshares between error rate nulling and error control based on maximum and minimum error thresholds. Error rate is visually sensed via a washout filter whose break frequency is set by the presence or absence of physical self-motion. It is shown why stick motion by itself does not provide a good estimate of workload, and how pilot compensation must be accounted for when using stick activity to compute workload. SCOPE’s range of application is substantially enhanced by two workload estimation techniques that use alternate information when the error signal being tracked is unknown. The first technique approximates pilot compensation from the vehicle dynamics which are used in the Crossover Model, and the second employs the vehicle output. Practical application of SCOPE’s workload estimation techniques is demonstrated using experimental data collected from a tracking task. Correlation between SCOPE’s workload estimate and pilot Bedford ratings was very high (0.93). Crossover frequency (which is needed for the Crossover Model) is estimated from the pilot data using the recently developed Equivalent Crossover System identification method. SCOPE results are compared with the Dynamic Interface Modeling and Simulation System Product Metric (DIMSS-PM) that has often been used to evaluate pilot workload.

Funding Sponsor

San José State University

Department

Psychology

Link to Full Text

Share

COinS