Abstract
Interruptions in the middle of a task have considerable costs. The objective of this study was to develop a system that postpones interruptions when they occur in periods of high workload. In Experiment 1, an air traffic control (ATC) simulator was presented with varying working memory demands. Pupil data were used to train a range of machine-learning classifiers to distinguish between high and low workload moments. The Gradient Boosted Tree (GBT) provided the best predictions. In Experiment 2, this classifier was used to develop a real-time interruption management system (IMS). The role of the IMS was to predict high and low workload and to postpone interruptions to the next low workload moment. To examine the IMS’s performance, its interruptions were compared to random interruptions. Results showed that the IMS successfully identified high and low workload moments with 76% accuracy, and postponed interruptions to the next low workload moment.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call