Performance Measures of Human Tracking Utilizing PID Modeling and a Closed Loop Error Metric.

摘要

Data from an antiaircraft artillery simulator is used to validate the dependence of tracking performance measured by the closed loop error and parameters obtained from a Proportional Integral Derivative (PID) model of the man. A comparison of three tracking teams is used as the data base with an azimuth and elevation operator composing each team. Although this problem is nonstationary, precognitive (a learned tracking task), two axis, and very difficult to model by conventional means, the PID model provides a simple approach to relate the error in the closed loop to the model parameters from each team. In order to provide adequate statistical estimates of these model parameters, identification techniques are applied to the same team (replications) tracking a similar arc tangent flyby trajectory. Statistical analysis was applied to the identified parameters between trackers to evaluate the credibility of this approach.