Driving behavior in Hazardous situations: The interplay between risk scenarios and dimensional emotions

Abstract

As the contradictions among the human-vehicle-environment elements in road traffic systems intensify, the resulting traffic safety issues are becoming increasingly severe. Emotions, as critical psychological factors influencing safe driving, directly impact drivers’ perceptions and judgments of surrounding information. Simultaneously, potential risks during the driving process can affect drivers’ decisions and driving behavior. On the basis of the definition of dimensional emotions, this study analyzes the impact of different valence, arousal, and risk level scenarios on drivers’ collision avoidance behavior. A total of 21 drivers aged 18 to 50 years, with driving experience ranging from 1 to 10 years, participated in a simulated driving experiment. This study employs the decision tree (DT) algorithm to define the quantitative relationship between the occurrence of different driving behaviors and their influencing factors. It predicts vehicle avoidance behavior in two emergency conditions to determine whether collision avoidance can be achieved. The results indicate that the model’s prediction accuracy improves when emotion and risk level information are combined, reaching 88.89 %. This is a 5.56 % improvement over the model using only emotion information and a 2.53 % improvement over the model using only risk level information. Compared with the impact of subjective emotional factors on driving behavior, risk scenarios exhibit a more stable trend in influencing driving behavior. Under the interaction of emotions and risk factors, the accuracy and generalization ability of the driving behavior prediction model based on the Decision Tree (DT) algorithm have been greatly improved. This research provides a theoretical basis for addressing adverse driving behavior caused by driver emotions and analyzing the factors influencing collision avoidance processes.