Ocular metrics for detecting attentional tunnelling

Abstract

This paper focuses on ocular measurement to detect the human operator’s particular state of “attentional tunnelling” during a robot supervisory task. After a survey of the existing ocular metrics, an innovative fixation detection algorithm is proposed. Then the metrics derived from the ocular parameters calculated by the algorithm are tested in a human-robot experiment. Among the metrics calculated, 3 of them appear to be able to statisticaly discrimintate the operators who faced attentional tunnelling.