Some modern control techniques for human operator modeling and identification

Abstract

Some recent results in quantitative modeling of the human operator are reviewed as well as a method for processing his input-output data. Modern control techniques are applied to the pilot operating in a dynamic compensatory tracking task. Trained pilots are known to behave in an optimal manner under fully stressed conditions. This has led to the development of an optimal control model of the human operator which takes his limitations into account. A rationale for determining the appropriate cost functions is determined. Experimental results indicate that the model can closely predict pilot performance in one- or two-axis tracking tasks. The model predicts pilot remnant (noise), error scores, and frequency response in both single-axis and multi-axis tracking tasks.