Following a remarkable increase in traffic efficiency and safety through the introduction of both passive and active safety technologies, automated vehicles are currently being introduced into road traffic with the intention to provide an even higher standard. The introduction of automation into other domains has shown that the anticipated benefits can also be accompanied by unexpected or novel problems. The step from assistance to automation does not simply mean the addition of more of the same but significantly alters the operator’s role and responsibilities, and the nature of task demands. Taking this into account, a driver-centered perspective is a key requirement for a successful introduction of automated vehicles.
Driver-centered design requires an understanding of the capabilities and vulnerabilities of both driver and driving automation system, as well as how they vary according to different tasks and situations. This thesis investigates the relevance of individual differences in the driver’s interaction with automation with a focus on take-over situations in five studies. A literature review, Article 1, identifies and classifies potential or established individual differences relevant for the interaction with a driving automation system. These individual differences are then positioned within human information processing to deduce their underlying causal mechanisms. The hypotheses about potential individual differences derived from this process are tested in four empirical, confirmatory driving simulator studies. In Article 2, the influence of the driver’s age on take-over performance in different traffic densities is investigated. Article 3 studies the ability to multitask sequentially in automated driving, i.e. taking over vehicle control while being engaged in a non-driving-related task. Article 4 investigates the influence of trust in automation on reliance and take-over performance. In the course of this study, a questionnaire to measure trust in automation was developed and evaluated, and is critically discussed. Article 5 investigates the monitoring task in partially automated driving with a focus on the driver’s ability to sustain attention and boredom proneness.
The results demonstrate that individual differences among drivers can crucially influence their interaction with automation, which in turn can have a critical impact on safety. An individual’s trust in automation influences how much a driver relied on an automated driving system and predicts whether a critical take-over situation will be successfully resolved. The ability to multitask sequentially predicts take-over time during engagement in a non-driving-related task. However, not every driver difference translates into an observable difference in the interaction with automation: Contrary to expectations, there is no evidence for an effect of age on take-over performance and no evidence for a relationship between the ability to sustain attention and the detection of a malfunction. In sum, the studies systematically revealed that individual differences known to be relevant in interaction with automated systems generally can also have a safety-relevant influence on a driver’s interaction with a driving automation system. The heterogeneity of the results underlines the importance of taking individual differences into account and highlights the relevance of a driver-centered perspective in automation design.