Since the COVID-19 pandemic, the number of traffic offenses and accidents related to aggression has increased, bringing back attention to the persistent problem of road rage for both industry and academia. Driven by practical use cases, a goal is the quantification of aggressive driving, as seen in telematics-based insurance premiums, the simulation of aggression in accident research, or the legal monitoring of aggressive driving behavior.

In this talk, a psychology-driven approach is introduced for modeling the pathway leading to aggressive driving. It is assumed that drivers are confronted with sources of frustration within the driving situation, which elicit negative emotions and culminate in an aggressive action when the individual’s emotional threshold is exceeded. Thus, frustrating sources act as contributors to aggression, based on which the prediction and classification of an aggressive action can be optimized. For instance, a short following distance is more likely to be classified as aggressive when preceded by a series of frustrating situations, such as red traffic lights or another vehicle overtaking on the right.

To investigate which emotional sources exist, when, where, and how frequently they occur, and which emotional value they hold, a real-world driving study was conducted, involving 34 participants, who tracked sources that triggered emotions in them. This talk presents the key sources of frustration identified in the study, and makes an initial attempt to reconstruct them using telematics data.