Support vector machine approaches to classifying operator functional state in human-machine system

The safety of a sophisticated human-machine system is closely related to the operator functional states (OFS) and the OFS is also associated with the physiological and mental state. The precise evaluation of OFS can realize an adaptive auxiliary system to assist operator decrease the accident risk. This paper has assessed and classified OFS based on a series of operator performance data, subjective evaluation data and electrophysiological signals. The classification model of OFS is based on the support vector machine (SVM). Analysis demonstrated that SVM can classify the OFS into three levels and the suitable feature selection contributes to increasing correct classification rate.