ABSTRACT
Cognitive distractions affect driver central cognitive processes. Visual distractions influence perceptual sensors and, to some extent, motor responses. Consequently, driver overt behaviors do not provide a sensitive or reliable basis for detecting driver cognitive distraction. This study demonstrated the value of indices of internal cognitive processes, specifically situation awareness measures, for cognitive distraction classification using a support vector machine (SVM) technique. In addition, the study establishes the importance of domain knowledge on the effects of distraction for promoting state classification accuracy. It also establishes the need to develop classifiers for different driving control modes.
