Abstract
The effects of an in-vehicle voice control system (VCS) on cognitive workload and driving performance were evaluated using a driving simulator study with 24 participants. Participants were asked to perform two types of in-vehicle tasks while driving: voice-command based radio and navigation tasks. The tasks were of two difficulty levels (easy, hard) with half of the tasks audio only and the other half audio with a visual display. Cognitive workload was measured using a tactile detection response task (TDRT) and a revised remote detection response task (RDRT). Driving performance was measured using standard deviation of lateral position (SDLP) and standard deviation of vehicle speed. An analysis of covariance for standard deviation of speed was used to examine the effects of VCS display with task duration as the covariate. Three separate mixed models were used to evaluate the variation in SDLP, response time, and misses given VCS tasks and detection response task (DRT) type. No significant differences were observed between TDRT and RDRT, suggesting both methods capture congruent measures of workload and that neither interfere with driving performance. Engagement with each hybrid VCS task was significantly associated with an increase in DRT misses above baseline driving. Additionally, hybrid display tasks had larger variations in vehicle speed and were completed more quickly than equivalent audio only tasks. Increasing values of task duration were associated with greater variation in lateral vehicle position. The results of this study suggest that design of VCSs should consider the modality, temporal components, and difficulty of tasks to reduce cognitive load.
Get full access to this article
View all access options for this article.