Projecting the planned trajectory of a Level–2 automated vehicle in the windshield: Effects on human drivers’ take–over response to silent failures

Abstract

To enhance the take–over performance by human drivers of Level–2 automated vehicles (AV), we developed a design concept that presents the AV's planned trajectory as augmented reality in the windshield. We hypothesized that, even when the AV does not release a take–over request before a potential crash (i.e., silent failure), the planned trajectory would allow the driver to foresee the crash and enhance the take–over performance. To test this hypothesis, we conducted a driving–simulator experiment where participants monitored the driving status of an AV with or without the planned trajectory in the context of silent failures. The results showed that, when the planned trajectory was projected in the windshield as if it were an augmented–reality display, the crash rate decreased by 10% and the take–over response time decreased by 825 ms compared to when the planned trajectory was not provided.