Operators' Monitoring Patterns and Fault Recovery in the Supervisory Control of a Semi-Automatic Process

Abstract

Although technological innovations have changed the role of operators from active participants to supervisors of semiautomatic processes, an understanding of the cognitive demands of supervisory control has not kept pace. In particular, little is known about when, and how well, operators might intervene and switch control from automatic to manual. This research addresses this issue by monitoring the information use and control actions of operators of a simulated semiautomatic pasteurization plant. The results of this experiment shows that individual differences in operators” monitoring patterns during the normal operation of the plant correspond to differences in their ability to mitigate the effects of faults. Specifically, an operator who controls the plant well during both normal and fault conditions tends to observe the plant frequently, integrating control actions with other control actions, and does not fixate on narrow sub-systems of the plant. On the other hand, an operator who performs poorly when exposed to faults tends to observe the plant less often, fails to integrate control actions, and fixates attention on a narrow subset of plant variables. Although all operators interacted with the plant using the same interface and automation, large individual differences in the operators” monitoring patterns, and the associated differences in performance suggest that individuals” attitudes, motivation, and training may play a critical role in the successful implementation of automation.