Due to the importance of individual mobility, the majority of people nowadays use a car for transportation. While doing so, it seems to be normal to perform additional, non-driving related tasks next to the main task of driving. Even though this task itself is complex and human cognitive resources are limited, serious accidents occur relatively seldom. Empirical studies suggest that in the context of driver-operated driving, drivers seem to have strategies at their disposal that enable them to guarantee a safe driving performance even in multitasking situations. They seem to have a target value in the driving performance, which they try to reach by constantly aligning the current level of driving performance to it. Furthermore, drivers seem to reduce the processing of non-driving related tasks when approaching a critical situation. In recent years, the development of automated driving functions has made considerable progress. This led to fundamental changes in the driver-vehicle-interaction. In automated driving (SAE level 3), the responsibility of the driving task is taken from drivers during normal traffic and they are permitted to engage themselves in other, non-driving related tasks. Due to this fact, the question arises whether and, if so, which role driver strategies might play in automated driving. The current research project concerns this subject. Based on the idea of a target value in the driving performance, thresholds of accepted driving performance values were defined in dependence on different individual and situational factors. Furthermore, it was investigated for different system configurations to what extent driver strategies are available in takeover situations from automated to non-automated driving. Finally, the results of these two steps were used to make recommendations for the design of future automated driving systems. (Author/publisher)
Abstract