Using eye tracking for live measures of workload in a refinery control room process monitoring task

Process monitoring within a refinery control room can be both daunting and listless. The workload of control room operators can fluctuate greatly over a short period of time. It would be ideal if systems could automatically detect these changes in workload so that both the computer system and those supervising are made aware of instances when an operator might be overloaded. While eye tracking has been shown to correlate well with workload, the process industry has yet to study this phenomenon. 48 control room operators from a major refinery participated within this experiment, where task complexity was varied within scenarios by controlling for the number of process deviations within a simulated crude process. It was found that both blink rate and pupil size could be used to track the cognitive load of operators over time. NASA TLX was also measured but it was not as sensitive as pupil size, as it did not capture a difference in scenario difficulty.