Better Together: 3D Anthropomorphic Assistant Avatars and Empathetic Voice Interventions on Emotion Regulation

Bofei Huang, Xinyue Gui, Chia-Ming Chang, Haoran Xie

Abstract: Despite advancements in driver-assisted systems, human emotions significantly influence driving safety and behavior. Previous research has highlighted the impact of emotions like anger and anxiety on driving, suggesting the need for effective emotional regulation strategies. Our approach leverages the visually engaging 3D assistant avatars and the calming effect of empathetic voice cues to enhance the driver’s emotional state, particularly in stressful scenarios such as traffic jams, encounters with uncivil driving behavior, and sudden pedestrian crossings. A video-based user study was conducted with 16 participants to evaluate the effectiveness of this combined approach. The study compared the proposed method against the baseline of empathetic voice alone across three specific driving situations. Using the Augmented Video testing platform, participants experienced simulated driving scenarios, followed by interventions from either the proposed or baseline method. Results from a two-item, 5-point Likert scale questionnaire indicate a preference for the combined method, highlighting its potential to improve driving safety and promote a more civilized driving experience. This research contributes to developing in-vehicle emotional regulation systems, offering significant insights into managing driver emotions to enhance overall road safety.