A Feature Selection Approach in Evaluation of Parameter Changes on the Human Operator Under Thermal Stress

Abstract
A feature selection approach is used lo define changes in the parameters (input-output characteristics) of the human operator involved in a closed loop tracking task subjected to thermal stress. By choosing a class of features which characterise the general shape of the human operator's transfer function, significant changes in these characteristics are found lo occur under the thermal stress condition. A F-Ratio Test was used with an Analysis of Variance to test the significance of the change in those features which describe the input-output characteristics of the human. The results of this experiment indicate that the effect of thermal stress (and the type of electrolyte used in the recovery period) will significantly change the input-output characteristics of the human operator. Therefore, the thermal stress condition is used to illustrate how this physiological stress condition manifests itself in the type of analysis performed here.