Driver experience and safety during manual intervention in a simulated automated vehicle: Influence of longer time margin allowed by connectivity

Deployment of personal automated vehicles is an on-going process. However, automated systems are expected to rely on human control and cooperation for a consequent amount of time. Thus, it is essential to (1) design carefully the interaction between the automated system and the driver and, (2) ensure proper trust and acceptance of the system. One way to do that is by providing drivers with longer time margins to prepare to intervene in driving in response to a takeover request. Extended sensing and prediction enabled by vehicle-to-infrastructure (V2I) connectivity might help to reach this goal. The experiment presented here aimed at investigating potential benefits of V2I on the interaction between the driver and the automated vehicle. In this study, carried out on a driving simulator, participants experienced two driving scenarios; one with an automated vehicle connected by V2I and one with a basic non-connected automated vehicle. Each scenario encompassed a takeover request caused by a roadwork zone. The vehicle connected with V2I allowed an anticipated takeover request. Those scenarios aimed at comparing (1) the safety of the driver intervention (2) driver's experience and mental workload during manual takeover and, (3) subjective evaluation of the system, with the connected and the basic automated vehicle. Results revealed a positive effect of V2I on the quality of driver's safety intervention in response to a takeover request and on the drivers’ subjective evaluation of the system and experience during the manual intervention.