Assessing driver frustration: Results from driving simulator studies

Abstract

Human beings often experience frustration when goal-directed actions are blocked. As driving is generally goal-directed (i.e., reaching a destination in a fast way) with a lot of obstacles (e.g., red lights, construction sites, slow lead vehicles), car drivers are especially vulnerable to frustration. Frustration in turn can lead to anger, which has extremely negative consequences on driving behavior, ranging from reduced situation awareness to risky driving, and even road rage. Therefore, it is desirable that future adaptive assistance systems can counteract a driver’s level of frustration, which at first requires a real-time recognition of this affective state. We will present two driving simulator studies that addressed which information can be used to assess spontaneously experienced frustration behind the wheel. In the first study, video recordings of the drivers’ faces were utilized to infer the level of frustration from the facial expressions. The results indicate that the activity of facial muscles in the mouth region can be used to classify drivers’ frustration from a neutral state. The second study evaluated whether a driver’s frustration can be assessed with functional near infrared spectroscopy of the brain. First results show that frustration (compared to a neutral state) elicited particular activation patterns in the lateral pre-frontal cortex. Finally, the presentation will conclude with a discussion of the acquired offline results and an outlook about the necessary steps towards an online recognition of spontaneously experienced frustration in the car based on multimodal signals