Alternative treatments of attention-sharing within the optimal control model

The effects of scanning (i.e. time-on-task) on manual control performance can be mimicked by a decrease in the operator´s signal/noise capabilities, provided scanning is rapid with respect to closed-loop system bandwidth. A simulation implementation of the optimal control pilot model was used to compare two treatments of attention-sharing and, for the scanning model, to explore the effects of scan cycle time. For the assumption of rapid scanning, model analysis showed that: (1) predicted tracking error is significantly degraded by reducing attention to the task, (2) the time-sharing (scanning) and capacity-sharing (signal/noise) treatments yield results that are significantly different, and (3) one can trade off between the time-sharing and capacity-sharing treatments in terms of the predicted performance/workload relationship. Increasing the cycle time of the scanning model yielded larger predicted tracking error scores, lower predicted pilot gain, and a substantially less Gaussian control behavior