Research on auditory and olfactory regulation methods for abnormal driver emotions based on EEG signals

In sudden and dangerous traffic situations, drivers are susceptible to abnormal emotional states, such as tension and anger, which can significantly increase safety risks while driving. Electroencephalography (EEG) signals, being an objective measure of emotional states, offer valuable insights for identifying and regulating these emotions.

This study collected EEG data from 54 drivers in a simulated driving environment, resulting in a total of 1,260 samples, and developed a recognition model for abnormal emotions-specifically tension and anger-based on the EEG signals. Time-frequency domain features, including mean, variance, skewness, kurtosis, root mean square, and power spectral density, were extracted and analyzed using classification algorithms such as Back Propagation Neural Networks (BPNN), K-Nearest Neighbors (KNN), and Support Vector Machines (SVM), enabling precise identification of varying levels of tension and anger. Furthermore, the study assessed the effects of music, fragrance, and their combined application on alleviating these abnormal emotional states.

Results indicated that music, fragrance, and their combination were related to a reduction in stress and anger across different severity levels, with subjective assessments correlating well with the objective EEG data. Notably, music regulation was found to be most effective for mild and moderate tension, reducing tension levels by 63.33% and 68.75%, respectively, whereas fragrance was more efficacious in high tension situations, achieving a 43% reduction. For anger, fragrance regulation proved more beneficial for mild and moderate anger (reducing anger by 66.67 and 73.75%, respectively), while music regulation was most effective in mitigating high anger levels, resulting in a 58% reduction. Additionally, an analysis of time-domain features utilizing Hjorth parameters revealed that the application of a single fragrance was most effective for alleviating tension, while a singular music regulation strategy demonstrated superior performance in calming anger.

The reliability of both the abnormal emotion recognition model and the emotion regulation assessment system was validated through the study. These findings contribute valuable scientific evidence for the management of drivers' emotions and suggest promising avenues for optimizing personalized emotional regulation strategies in the future.