Abstract
The automotive industry has recently made first steps towards implementation of automated driving, by introducing lateral control as addition to longitudinal control (i.e. ACC). This automated control is allowed during specific situations within existing infrastructure (e.g. motorway cruising). During these circumstances, the role of the driver changes from actively operating the vehicle to supervising the system. Due to being placed remote from the control-loop, vigilance and overreliance, performing supervisory tasks is something humans are typically not very good at. For this reason, Human Factors experts have often raised concerns about the implementation of semi-automation. Nonetheless, we observed that recommendations on how to design appropriate interaction between driver and automation, are rather scarce. Therefore, we reviewed Human Factors’ literature and three existing interface examples to retrieve recommendations on desired driver-vehicle interaction. The most important design considerations were: (a) Avoid mode confusion by informing the driver appropriately about system state; (b) Support awareness of the system’s operational envelope, i.e. helping drivers understand the boundary limits within which the automation is able to operate; (c) Provide instructions with respect to the required role of the driver. From the examples reviewed, we concluded that interfaces representing a road situation graphically, while depicting elements relevant for system functioning (e.g. detection of road lines and/or target vehicle), provide effective solutions to support driver’s awareness of the system’s operational envelope. Nonetheless, we observed that there is at this moment no univocal understanding of what kind of interface mechanism works overall best. No consensus has been reached on appropriate ways to communicate mode-changes and to effectively instruct drivers when a role-change is required (i.e. retrieving control). Because confusion might easily occur when needing to interpret mode-information in time-critical situations and revealing the associated driver’s role, we strongly recommend more focus on drivers’ instructions in the development of future interfaces.
KEYWORDS – semi-automated driving, human machine interface, supervisory control, retrieving control, system awareness, mode changes