Abstract
This study aims to clarify appropriate setting of dual task methodology to grasp the impacts of in-vehicle voice-command interfaces on actual driving behavior. Steering wheel tracking, accelerator pedal tracking, and visual detection response tasks were examined as the candidates of performance measurement task that would be carried out with the target assessed tasks. As the results of experiments, it was found that pedal tracking task was superior as the performance measurement task in terms of the sensitivities among target task differences and the correlations with measures of vehicle lateral displacement and headway distance in actual driving.