<p>This study investigates the feasibility of detecting driver emotions using electrodermal activity (EDA). In the first experiment, four emotion induction methods—film watching, writing passages, interview, and simulated driving —were evaluated. In the first experiment, thirty-nine participants took part, and film watching emerged as the most effective method to induce anger in terms of valence and arousal. In the second experiment, the film-watching method was used to induce four emotions—anger, surprise, happiness, and boredom—and corresponding EDA responses were analyzed using Russell’s circumplex model of affect. Forty participants participated in second experiment. The results showed a significant increase in tonic EDA levels during high-arousal emotions (anger, surprise, and happiness), whereas the phasic-max value significantly decreased only during surprise. These findings suggest that EDA predominantly reflects emotional arousal rather than valence, highlighting its limitations for fine-grained emotion recognition when used alone. To enhance accuracy and robustness, future studies should incorporate additional physiological signals, such as electromyography, along with contextual and psychological factors.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Validating Emotion Induction Techniques and Exploring EDA-Based Driver Emotion Detection

  • Bogyu Choi,
  • Youngbin Kwon,
  • Young Dal Oh,
  • Ji Hyun Yang

摘要

This study investigates the feasibility of detecting driver emotions using electrodermal activity (EDA). In the first experiment, four emotion induction methods—film watching, writing passages, interview, and simulated driving —were evaluated. In the first experiment, thirty-nine participants took part, and film watching emerged as the most effective method to induce anger in terms of valence and arousal. In the second experiment, the film-watching method was used to induce four emotions—anger, surprise, happiness, and boredom—and corresponding EDA responses were analyzed using Russell’s circumplex model of affect. Forty participants participated in second experiment. The results showed a significant increase in tonic EDA levels during high-arousal emotions (anger, surprise, and happiness), whereas the phasic-max value significantly decreased only during surprise. These findings suggest that EDA predominantly reflects emotional arousal rather than valence, highlighting its limitations for fine-grained emotion recognition when used alone. To enhance accuracy and robustness, future studies should incorporate additional physiological signals, such as electromyography, along with contextual and psychological factors.