Effects of training and display content on Level 2 driving automation interface usability

Mueller, Alexandra S. / Cicchino, Jessica B. / Singer, Jeremiah / Jenness, James W.
Insurance Institute for Highway Safety
June 2019

Introduction: Advanced driver assistance systems have the potential to improve safety, but as they become increasingly sophisticated there is a growing risk for drivers to misunderstand their functionalities and limitations. Level 2 driving automation features, such as adaptive cruise control (ACC) combined with lane centering, primarily communicate their operating statuses to the driver through the instrument cluster. It remains an open question how interface-specific training and display content influence a driver’s use and comprehension of Level 2 automation in production vehicles.
Methods: A total of 80 participants viewed videos recorded from the driver’s point of view under a variety of driving conditions with level 2 automation activity displayed in the instrument cluster of a model year 2017 Mercedes-Benz E-Class. Half of the sample viewed one of two instrument cluster layouts (simple or complex), and half received a brief orientation to the interface of their experimental group prior to the experiment. Participants viewed videos recorded from the driver’s point of view under a variety of driving conditions with Level 2 automation activity displayed in the instrument cluster. After each video they were asked about the scenario they had just seen. We then examined what information in the instrument cluster participants used to identify Level 2 automation activity and their perceived usability of the displays.
Results: Training improved the ability to detect when lane centering was temporarily inactive and improved understanding of why the system was inactive. Neither training nor instrument cluster content affected the ability to identify when ACC had adjusted the vehicle’s speed or detected a vehicle ahead, which all participants were highly accurate at detecting regardless of condition. The experimental factors also had no effect on the identification of when ACC initially did not detect a lead vehicle and the understanding of why it had not detected it, on which performance was universally poor. Both factors, however, influenced which sources of information in the display participants relied on to determine Level 2 automation activity. In turn, accuracy of system activity detection improved when participants relied on the correct sources of information. Training, but not instrument cluster content, also influenced the perceived usability of lane centering but not of ACC.
Discussion: Basic training improves detection of some system notifications that potentially require further driver action, but not of those that display persistent status information, where understanding without training is high. Even so, training does not result in full understanding of all system notifications or functional limitations, which reinforces the need for vehicle interfaces to saliently communicate pertinent system behavior and its limitations intuitively to naïve users.