Cognitive demands in automation

As the nature of a task changes with increasing automation, so do the cognitive demands on the human operator. The present study examines the implications of such changes on subjective workload and automation design issues. A variety of tasks (continuous flight control, discrete target acquisition, and decision making) was used to tax the different cognitive systems. Each task could be performed alone or in combination with other tasks. Performance and subjective workload ratings, obtained by three workload scales, were examined as a function of the level of automation. Results demonstrated the usefulness of the multiple resource approach to task analysis. All three workload scales were found to be sensitive to the task demands. The Bedford scale was particularly impressive in distinguishing the different types of resources in demand. Both the multiple resource model's performance predictions and the subjective workload ratings were shown to be invaluable for decisions on how and what to automate.