Adaptive Automation of a Dynamic Control Task Based on Secondary Task Workload Measurement

Adaptive automation (AA) has been proposed as a method for regulating human workload and abating out-of-the-loop performance problems in complex systems control. The majority of AA or adaptive manual control studies, to this point, have facilitated control allocations using either preprogrammed schemes based on desired system performance, comparisons of human performance with established goals, or psychophysical variable monitoring to represent workload levels and determine appropriate control allocations for moderating workload. This study further explored the psychophysical assessment approach by using a secondary task measure of workload to facilitate control allocations in a complex primary task. An experiment was conducted in which participants performed a secondary gauge-monitoring task along with a simulated radar monitoring and target elimination task. Differences in single-secondary task performance and performance observed while also functioning in the primary task were used to direct operator-managed shifts of the level of automation in the primary task. Shifts between manual control and partial automation were mandated for 1 group of participants and merely suggested for another. Changes in the workload measure were observed in associated control allocations. In addition, subjective ratings of workload were recorded. Results revealed significantly improved manual, primary task performance and enhanced secondary task monitoring under automation for the mandated AA group, with the opposite results occurring for nonmandated AA participants. Participant workload marginally exceeded an objectively established criteria by using the secondary task measure to direct AA of the primary task.