2024 Volume 2024 Issue 12 Article ID: JRJ20241202
As the deployment of conditional Automated Driving Systems (ADS) becomes increasingly prevalent, particularly in urban settings, there is a potential that users may not comply with the system's behavior and may manually override it based on their own judgment, even without a Request to Intervene (RtI) from the system. Such driver-initiated interventions could pose a risk to the safety of automated driving. This study simulated a scenario where a conditionally automated vehicle overtook a parked vehicle on a two-lane street with oncoming traffic to examine the circumstances surrounding driver-initiated interventions. The results revealed that all instances of manual override were executed by pressing the brake pedal and seizing the steering wheel while the automated vehicle was approaching the parked vehicle in close proximity to the oncoming traffic traveling in the opposite direction. Out of the 23 participants, 12 expressed confidence in the system's decision-making and performance, 9 felt generally safe but experienced specific apprehensions and discomfort about the system's operation, and 2 expressed concerns about safety issues related to the system's perception and proximity to surrounding vehicles. These findings offer valuable insights into the factors and patterns of driver intervention and can inform the design of ADS.