It is possible that eye fixations reflect the processing state of the driver and may provide some insight into the driver's cognitive model of the situation. The goal of this project is to characterise how eye fixation patterns change as the driving situation changes. Eye movements are recorded while driving in a simulator under controlled visual and road conditions. In the first task (driving on an open road), the eye movements are modelled with Markov matrices and identify two scanning patterns. In the second task (following a lead car) a new Markov matrix is derived in which a new pattern of movement has been added to the existing two patterns. The results support the idea that Markov models of eye movement behaviour in simple situations can be linearly combined to predict behaviour in complex tasks.
Samenvatting