ABSTRACT
Psychophysiological data has been used to detcrmine the mcntal workload state of operators. Thi s has bccn successfully made use of to drive adaptive aiding. In order to be employed in workstations the psychophysiological data must consistently providc accurate information on the state of the opcrator. Using a complex task and an artificial neural nctwork classifier highly accurate classification of operator state was found within one day. However, the accuracy of the estimates decreased when data from different days were used to test the classifier. Adjustments to opcrator statc classification procedures may have to be made to accommodate the day-to-day variabi lity of thc psychophysiological measures.
