A Novel Method of Emergency Situation Detection for a Brain-Controlled Vehicle by Combining EEG Signals with Surrounding Information

In this paper, to address the safety of brain-controlled vehicles under emergency situations, we propose a novel method of emergency situation detection by fusing driver electroencephalography (EEG) signals with surrounding information. We first build a novel EEG-based detection model of driver emergency braking intention. We then recognize emergency situations by fusing the result of the proposed EEG-based intention detection model with that of the obstacle detection model based on surrounding information. The real-time detection system of driver emergency braking intention is implemented on an embedded system, and the driver-and-hardware-in-the-loop-experiment of the proposed detection method of emergency situations is performed. Experimental results show that the proposed method can detect emergency situations with the system accuracy of 94.89%, false alarm rate of 0.05%, and response time of 540 ms. This paper has important values in the future development of brain-controlled vehicles, human-centric advanced driver assistant systems, and self-driving vehicles and opens a new avenue on how cognitive neuroscience may be applied to human-machine integration.