The Hazard Potential of Non-Driving-Related Tasks in Conditionally Automated Driving

Today, humans and machines successfully interact in a multitude of scenarios. Facilitated by advancements in artificial intelligence, increasing driving automation may allow drivers to focus on non-driving-related tasks (NDRTs) during the automated ride. However, conditionally automated driving as a transitional state between human-operated driving and fully automated driving requires drivers to take over control of the vehicle whenever requested. Thus, the productive use of driving time might come at the cost of increased traffic safety risks due to insufficient and insecure human-vehicle interaction. This study aims to explore the take-over performance and risk potential of different NDRTs (auditory task, visual task on regular display, visual task with mixed reality hardware) while driving. Our study indicates the hazard potential of visual vs. auditory distraction and multitasking vs. sequential tasking. Our findings contribute to understanding what influences the acceptance and adoption of automated driving and inform the design of safe vehicle-human take-overs