关键词: 交通运输工程； 眼动与脑电特征；多特征融合；疲劳驾驶；检测模型

Abstract: Fatigue driving is the main cause of traffic accidents. EEG or eye movement features are commonly used to detect driving fatigue. Compared with eye movement features, EEG features have higher detection rate, but the stability is not as good as that of eye movement features. In order to obtain high and stable detection rate, the wavelet entropy function was firstly used to extract the EEG features from the EEG electrodes CP4, TP8, T5, P3, Pz, P4, T6, O1, Oz and O2. And then EEG features and eye saccade length were fused, and KNN algorithm was used to establish a fatigue detection model based on fusion features. Finally, the detection model based on fusion features was compared with the detection model based on EEG features. The results show that: the detection mean values of the two detection models are not much different. but the detection effect based on fusion features is more stable than that based on EEG features.However, in terms of detection stability, the model based on fusion features is more stable than the model based on EEG features

Key words: traffic and transportation engineering; eye movement and EEG features; multi-feature fusion; fatigue driving; detection model