A discriminating method of driving anger based on sample entropy of EEG and BVP

“Road rage” has been a concern to traffic safety and management authorities. In order to provide an effective detection of angry driving behavior, it is necessary to provide a method for discriminating angry driving from normal. Thirty professional drivers were recruited for conducting on-road experiments in Wuhan, China. The drivers were required to finish the experiment within 110 minutes and their anger was induced by various stimulating events, e.g., jaywalking and weaving/cut-into of other vehicles. The electroencephalography (EEG) and blood volume pulse (BVP) signals of drivers were collected in the normal and angry states from the field experiments. Statistical analysis shows that the sample entropy of EEG and BVP signals was viable to be used as the index for identifying angry driving. Based on the obtained EEG and BVP sample entropy, a receiver operating characteristic (ROC) curve analysis was introduced to determine the discriminating threshold of driving anger. The results indicate that, when the EEG sample entropy is between (0.2717, 0.6867) and the BVP sample entropy is between (0.4816, 0.7056), the driver is in the transitional period, which means that the driver can become angry easily when facing the stimulating events. When the EEG sample entropy is smaller than 0.5817 and the BVP sample entropy is bigger than 0.6037, the driver is likely to be in an angry state, with an average accuracy of 80.41%. Therefore, it appears reasonable to use the EEG sample entropy of 0.5817 and the BVP sample entropy of 0.6037 as the threshold for identify driving anger. It is believed that the study results can provide a theoretical foundation for developing EEG and BVP sample entropy-based driving anger recognition and warning devices.