Comparison of Different Combinations of Ultra-Short-Term HRV Features for Identifying Driving Anger

Zhibin Zhu,Xuanyi Wang,Jingwen Feng,Yifei Xu,Wanlin Chen,Jing Zheng,Shulin Chen,Hang Chen

DOI: https://doi.org/10.1109/icicsp59554.2023.10390606

2023-01-01

Abstract:Objective: Anger is one of the basic negative emotions, and scenes such as road rage can pose a threat to public safety. The expression of anger and the characteristics of heart rate variability (HRV) are closely related to the autonomic nervous system. HRV features can be analyzed using ultra-short-term R-R interval data; Therefore, this article aimed to compare the effectiveness of different combinations of HRV features in identifying driving anger. Method: Anger and calmness were elicited using video clips during electrocardiography (ECG). 67 HRV features were from extracted and remove individual differences. The best 10 features were selected by using Mann-Whitney U-test. Then, Support Vector Machine (SVM) was used to calculate the accuracy and compare these combinations. Results: as the number of features in the combinations increased, the identification effect became better, but the improvement decreased. After the number of features exceeded 5, there was basically no improvement. Meanwhile, the gap between the combination and combination also decreased. Although the overall effect might not be as good when there were single or few features in them, certain specific feature combinations had good results, such as the combination of four features $(\text{SDNN}+\text{RR}- \text{Mean}+\text{Prc} 20\text{NN}+\text{pNN} 20)$ which was already one of the best combinations. Significance: This study can provide important support and reference for the identification and intervention of “road rage” and other scenarios.