How Does Instructed Knowledge Influence Drivers’ Decision-Making in Conditional Driving Automation?

This study aimed to reveal how instructed knowledge influences human drivers to make decisions in complex driving-task traffic environments when they use conditional driving automation (CDA). A driving simulator experiment was conducted to collect data in resuming car control from automation; 54 individuals (mean age = 40.9 ± 16.8 years old) participated. According to the experimental results, imparting knowledge of the request to intervene (RtI) helped to reduce the reaction time (RT) to the RtI to 1.22 s on average, but also caused 20% of drivers to resume vehicle control who failed to choose the correct route in terms of the issued route guidance. This ratio was significantly higher than the 3% of drivers who were unfamiliar with the RtI, but were only instructed on the necessity of taking over car control due to functional limitations. On the other hand, 28% of those drivers resumed control for responding to the guidance before the system limitation occurred. Further, subjective assessment suggests that the drivers tended to over-rely on their instructed knowledge, and were prone to overlooking other important details when the knowledge was restricted to the specified concept. Conveying knowledge appropriately not only contributes to driver interventions, but also greatly impacts other kinds of decision-making performance.