Research Question/Objective: Current advanced driver assistant systems combine the strengths of a human driver with the benefits of technical advancements. By raising the vehicle automation level, new human factors challenges emerge. Considering level 2 automation, where the driver is required to continuously monitor the system and remains responsible for vehicle safety, automation effects like overtrust and underload emerge and low vigilance and attention levels could impair driver performance. When reaching level 3 automation, the driver will still be needed as a fall back level for the automated system. Here, various automation effects could impair driver take-over performance and thereby controllability of the overall system, which is a combination of the system’s reliability and driver’s availability. To ensure the safety of the automated function, in-depth knowledge of the driver’s current state, and hence driver’s availability is essential. Moreover, both new standardized evaluation approaches and a common comprehension of the safety relevant parameters are necessary. In order to gain a better understanding of drivers in take-over situations, fatigue is examined in a driving simulator.
Methods and Data Sources: The paper summarizes the different existing methods to assess driver state and controllability of level 3 systems, and how aspects such as fatigue influence the driver within take-over situations. In a driving simulator study different fatigue levels were established by means of slight sleep deprivation combined with hypovigilance and rated on the Karolinska Sleepiness Scale. Driver performance was assessed in regard to timing and quality aspects of the take-over.
Results: The results indicate a correlation between fatigue and drivers’ take-over performance and proves the validity of the applied fatigue measures.
Discussion and Limitations: Fatigue was investigated in the driving simulator which may have limited validity. Fatigue presents particular challenges to the experimental setup, as it is difficult to be established artificially.
Conclusion and relevance to session submitted: The proposed paper examines human performance in a highly automated driving situation under the influence of different fatigue levels which helps to assess the safety of future automated vehicles.