A Probabilistic Model of Taking-Over Control from Semi-autonomous Vehicles

Abstract

In automated driving scenarios with semi-autonomous vehicles of today and the near future, a request for taking-over control may at any time be communicated to a human driver. Empirical studies exist of, e.g., the effects of different modalities used in this human-machine interaction, and of drowsiness of drivers involved, depending on manual or automated driving. However, there is no model available yet of how the decision-making time in the course of taking-over control from the automated vehicle depends on the driver’s drowsiness. Hence, we present a probabilistic model of the decision-making time as influenced by certain properties characterizing the driver’s drowsiness according to previous work.