Employing Eye Trackers to Reduce Nuisance Alarms

Area of Science:

• Human-Computer Interaction
• Cognitive Engineering
• Industrial Process Control

Background:

• Alarms that are anticipated by operators lose informational value and become nuisances.
• Reducing alarm salience when anticipation is detected could mitigate nuisance alarms.
• Eye tracking offers a potential method for inferring operator anticipation of alarms.

Purpose of the Study:

• To investigate the use of eye-tracking-based gaze-based acknowledgement (GBA) methods to estimate operator alarm anticipation.
• To evaluate the effect of reducing alarm salience based on GBA on control task performance.
• To determine if different process monitoring scenarios influence GBA method utilization.

Main Methods:

• Recruited 24 engineering students to predict alarms and control a process parameter.
• Implemented three GBA methods that reduced alarm salience upon detecting potential anticipation.
• Assessed performance across three scenario types: parameter near threshold, trending, or fluctuating.

Main Results:

• No significant improvement in control task performance was observed with any GBA methods.
• The type of process monitoring scenario significantly affected the utilization of different GBA methods.
• Alarm prediction accuracy was higher with gaze-based acknowledgements compared to standard methods (OR = 3.877, p < 0.001).

Conclusions:

• Gaze-based acknowledgement methods show potential for improving alarm prediction accuracy in process control.
• While not improving task performance in this study, GBA methods were preferred by participants over standard alarms.
• Further research is needed to enhance the validity of eye tracking for inferring operator alarm anticipation.