Towards a Cognitive Load Aware Multimodal Dialogue Framework for the Automotive Domain

A significant aim in developing multimodal HCI for the automotive domain is to keep driver´s distraction low. However, the measurement of the cognitive load is difficult and inaccurate but an approach to predict the effect of dialogue and presentation strategies on this is promising. In this paper we discuss cognitive load in theory and related work, and identify dialogue system components that play a role for monitoring and reducing driver´s distraction. Subsequently we introduce a dialogue system framework architecture that supports cognitive load prediction and situation-dependent decision making & manipulation of the HCI.