Development of a Model for Human Operator Learning in Continuous Estimation and Control Tasks.
Abstract
This research was directed toward the development of an analytic tool for the design of training procedures and the assessment of trainee performance in the kinds of monitoring, decision, and control tasks required for flight management. Manual control data obtained in previous AFAMRL Laboratory studies was analyzed with regard to learning behavior. This analysis consisted of three steps: model analysis with the optimal control pilot model (OCM) to determine the relations between stages of training and independent 'pilot-related' model parameters; tests of some hypotheses concerning the underlying effects of training on control-strategy development; and preliminary analysis to explore relationships between the perceptual cueing environment and the pilot's internalized representation ('internal model') of the task situation. The results of the analysis suggest that continued practice on the tracking task leads to a more precise, consistent, and linear (i.e., less 'noisy') type of response behavior, and to an improved internal model. Analytic results further suggest that, if the OCM is modified to account for the pilot's ability to construct his internal model, the model should be able to predict the effects of the task structure (including plant dynamics, input spectra, and cueing environment) on the rate at which (and degree to which) the human operator develops his estimation and control strategies.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1983
- Accession Number
- ADA140320
Entities
People
- W. H. Levison
Organizations
- BBN Technologies