A Fatigue Driving Detection System Based On Eye Detection And Eeg Detection

The Chinese automobile industry develop quickly in the past decade.Nowadays, the industry is in a golden period. However, the sudden increasing number of cars and drivers has brought great pressure to the transportation.Traffic accident rate keeps the top of the world for many years. In all traffic accidents, the ones caused by driver fatigue are up to twenty percent ratio. Thus the driver fatigue detection technology has practical significance to prevent the fatigue driving and to improve the safety of the transportation.In this dissertation, a novel feature extraction algorithm called Extend Nearest Feature Line Space and a fatigue driving detection system based on eye detection and EEG detection are proposed. The proposed feature extraction algorithm which applied for face recognition of the authentication module will extend the prototype sample set and then uses the extended sample set for feature extraction. The proposed system fusions two kinds of fatigue detection methods which work in coordination. The system also uses the driver’s personal physiological parameters to improve detection accuracy and to reduce the probability of false positives. The system mainly consists of four modules: brain wave detection module, the human eye detection module, the authentication module, and the information fusion module. Brain wave detection module is designed for collecting and extracting the brain wave signal. The eye detection module acquires the image signals through the camera and real-time analysis of the state of the human eye. Authentication module makes use of the face recognition method to authenticate the identity of the driver, and modify the parameters with the certified results. The information fusion module obtains information from the brain wave detection module and the human eye detection module, and combines the information with the parameters given in the authentication module, and then provides the ultimate fatigue determination.In driving simulation experiments, the performance of each module of the system has been validated. The personalized parameter system can achieve a higher detection accuracy and a lower false rate in the comparison with the average parameter system. Compared to the other similar systems, the proposed system not only achieves a high detection accuracy, but also show the simple architecture, low cost and strong anti-interference capacity.