Operator functional state assessment for adaptive automation implementation

Abstract

Mission success in military operations depends upon optimal functioning of all system components, including the human operator. The cognitive demands of current systems can exceed the capabilities of the human operator. In some situations, such as Unmanned Combat Air Vehicle (UCAV) operations, one operator may be required to supervise several vehicles simultaneously. The functional state of the human operator is not currently considered in the overall system assessment. It has been assumed that the operator could “manage” any situation given a well designed system. However, with the requirement to monitor and remotely monitor several vehicles simultaneously during combat comes the possibility of cognitive overload. This increases the probability of committing errors. We have developed on-line measures of operator functional state using psychophysiological measures. These measures provide estimates of how well an operator can deal with the current task demands. When the operator is cognitively overloaded the system may be able to implement adaptive aiding procedures. This will reduce the task demands on the operator thereby improving mission success. We have demonstrated correct assessment of the functional state of the operator with accuracies of 95% or better. Psychophysiological measures were used with classifiers such as artificial neural networks. In one study, adaptive aiding was applied when the classifier determined operator overload. The aiding resulted in significantly improved performance.