Maintaining attention with enhanced human-machine interfaces in conditionally automated driving

Vehicles are becoming increasingly automated. Recently, vehicles equipped with conditional automation have been approved for operation on certain roads, allowing the operator to engage in non-driving related tasks (i.e., texting, reading; NDRT). However, the operator must be prepared to take back control when the automated system can no longer perform properly, as indicated by a request to intervene (RTI).

Three sets of research studies evaluated different approaches to help the operator reorient their attention to the driving task when a takeover is required. The first set of studies investigated how blanking the NDRT could help the operator reorient with the driving task faster when compared to the NDRT remaining on. The second study evaluated how informing the operator of the automation’s confidence in its abilities via the human-machine interface (HMI) affected the transition of control. Finally, the third set of studies evaluated how the timing of the transition window and modality of the RTI influenced the transition of control.

Results found that participants were faster to look back to the roadway after an RTI under the following conditions: the NDRT was removed before the RTI; the automated system displayed its confidence in its abilities via the HMI paired with an auditory chime; the RTI was coupled with a brake pulse; and the transition window was 4s. Results found that participants were faster to disengage the automated system after an RTI under the following conditions: the NDRT was removed before the RTI; the RTI was coupled with a brake pulse; and the transition window was 4s.

With conditional automation currently in production stages, it is critical that the technologies are designed to best support the operator in reorienting to the driving environment and regaining situation awareness before taking control.