Driver readiness model for regulating the transfer from automation to human control

In the collaborative driving scenario of truck platooning, the first car is driven by its chauffeur and the next cars follow automatically via a so-called 'virtual tow-bar'. The chauffeurs of the following cars do not drive 'in the towbar mode', but need to be able to take back control in foreseen and unforeseen conditions. It is crucial that this transfer of control only takes place when the chauffeur is ready for it. This paper presents a Driver Readiness (DR) ontological model that specifies the core factors, with their relationships, of a chauffeur's current and near-future readiness for taking back the control of driving. A first model was derived from a literature study and an analysis of truck driving data, which was refined subsequently based on an expert review. This DR model distinguishes (a) current and required states for the physical (hand, feet, head, and seating position) and mental readiness (attention and situation awareness), (b) agents (human and machine actor), (c) policies for agent behaviors, and (d) states of the vehicle and its environment. It provides the knowledge base of a Control Transfer Support (CTS) agent that assesses the current and predicted chauffeur state and guides the transition of control in an adaptive and personalized manner. The DR model will be fed by information from the network and in-car sensors. The behaviors of the CTS agent will be generated and constrained by the instantiated policies, providing an important step towards a safe transfer of control from automation to human driver. © 2017 ACM. ACM SIGAI; ACM SIGCH