Implementing dynamic changes in automation support using ocular-based metrics of mental workload: a laboratory study.

Adaptive Automation has been often invoked as a remedy to indiscriminate introduction of automation support. However, this form of automation is difficult to implement without a sensitive and reliable index of the Operator Functional State. In a series of studies we have showed the usefulness of the distribution of eye fixations as an index of mental workload to be used as a trigger of automation. Particularly, the distribution pattern was found to be sensitive to taskload variations and types, thus making it very appealing for designing adaptive systems. This approach seems to be valid and reliable, but a necessary step in this research program would be testing the effectiveness of automation driven by fixation distribution and its capability in reducing the workload. The present study is a first attempt to carry out such validation.