Modeling human performance in a time-varying anti-aircraft tracking loop

Optimal control and human response theory are combined to develop a computerized, predictive model of the input-output tracking response of the human gunner in an antiaircraft artillery loop. The model includes representations for various human limitations, such as time-delay and randomness, that act to degrade system performance. The model extends earlier work in manual control by considering rapidly varying system dynamics, and arbitrary nonstochastic input disturbances. Model predictions of tracking error covariance are compared with human tracking data from Army conducted experiments that simulated the Vulcan Air Defense System (VADS). The agreements are excellent for both elevation and azimuth axis tracking, over a range of target passes from easy to difficult. The predictions are obtained using a single set of man-model input parameters that are typical of human response limitations. It is concluded that the modeling approach can be used with confidence in a system analysis study of weapons effectiveness.