Incorporating different phases of situation awareness information into the design of advanced takeover requests for urgent takeover scenarios

Advanced automated driving systems allow drivers to engage temporarily in non-driving-related tasks (NDRTs), such as sending messages or playing games. However, drivers still play a critical role in manually taking over the vehicle when the systems fail, necessitating prompt situation awareness development (SA). To accelerate SA development under strict time constraints, this study incorporated the concept of monitoring requests (MRs) into an advanced takeover request (TOR) design, offering three distinct SA-focused phases: Level 1 redirects attention to potential dangers; Level 2 facilitates the interpretation of cues, such as changes in relative velocity and remaining distance; and Level 3 projects future events, such as the time to collision (TTC). Moreover, two common takeover scenarios were considered to obtain insightful guidelines for creating tailored TOR designs. This study used a driving simulator and the Wizard of Oz method to simulate autonomous driving, collecting data on takeover reactions, vehicle movements, and subjective questionnaires from 32 participants. The statistical analyses indicated that TOR design highlights advanced levels of SA, particularly Level 3 SA, for future projection, which benefits overall SA quality and increases minimum TTC, enhancing safety margins during urgent takeovers. Furthermore, these improvements in the takeover were accompanied by increased changes in posttakeover vehicle control. These patterns suggest more active post-takeover control rather than a hasty response. Lastly, an advanced TOR can be tailored to specific takeover scenarios, with dynamic scenarios benefiting from Level 3 TOR design and static scenarios benefiting from Level 2 TOR design.