ABSTRACT
The wholesale use of automation has been associated with a number of potential human performance problems, including: increased monitoring and supervisory control demands; miscalibrated trust in automation; manual skills degradation; and cognitive overload. Adaptive automation refers to a system capable of dynamic, workload-triggered reallocations of task responsibility between human and machine. Although preliminary empirical work has provided evidence that adaptive automation might benefit both monitoring performance and manual skills retention, such evidence has generally been collected in the context of control and monitoring tasks. Pupil diameter was roughly equivalent under high traffic for the fixed manual and fixed automation schedules, perhaps an indication of a ceiling effect. Physiological data suggest that mental workload can be managed through the use of dynamic task allocation shifts. The ability of automation to provide workload reductions might have to be diluted, in the name of task involvement and keeping the human operator actively ‘in-the-loop’.
