Driver Training Research and Guidelines for Automated Vehicle Technology

The advent of advanced driver-assistance systems presents the opportunity to significantly improve transportation safety. Complex sensor-based systems within vehicles can take responsibility for tasks typically performed by drivers, thus reducing driver-related error as a source of crashes. While there may be a reduction in driver errors, these systems fundamentally change the driving task from manual control to supervisory control. A significant challenge, given this fundamental change in the driving task, is that there are no established methods to train drivers on the use of these systems, which may be counterproductive to safety improvements. The aim of the project was to develop training protocol guidelines that could be used by advanced driver-assistance system trainers to optimize driving safety. The guidelines were developed based on the results of three activities that included the development of a taxonomy of the knowledge and skills necessary to operate advanced driver-assistance systems, a driving simulator study that examined the effectiveness of traditional training protocols, and a test track study that examined the efficacy of a vehicle-based training protocol. Results of both studies suggest that differing training protocols are most beneficial in terms of driver cognitive load and visual scanning as opposed to short-term changes in performance.