摘要
疲劳驾驶是引发交通事故的重要原因之一。为有效检测驾驶员疲劳状态,对基于深度学习的驾驶员脑电信号疲劳检测方法进行研究,提出一种基于栈式自编码的深度学习框架,检测驾驶员疲劳状态。首先,对脑电信号进行预处理,通过离散傅里叶变化分段提取疲劳特征;然后,设计栈式自编码神经网络,检测疲劳状态,并分析不同特征对疲劳检测结果的影响。实验结果表明了栈式自编码对驾驶员脑电信号疲劳检测的有效性。该研究对驾驶员疲劳检测系统的开发具有重要意义。
Fatigue driving is one of the important causes of traffic accidents.In order to effectively detect the driver’s fatigue state,a deep learning-based driver’s EEG signal fatigue detection method was studied,and a Stacked Auto-encoder deep learning framework was proposed to detect the driver’s fatigue state.Firstly,the EEG signal is preprocessed,and the fatigue features are extracted by discrete Fourier transform segments.Then,a stacked auto-encoder neural network is designed to detect the fatigue state and analyze the influence of different features on the fatigue detection results.The experimental results show the effectiveness of stacked auto-encoder for fatigue detection of drivers’EEG signals.This research has important significance for the development of driver fatigue detection system.
作者
石锦璇
王昆
SHI Jinxuan;WANG Kun(College of Physics and Electronics Engineering,Shanxi University,Taiyuan 030006,China)
出处
《测试技术学报》
2023年第2期140-145,共6页
Journal of Test and Measurement Technology
基金
2021年度山西省基础研究计划(自由探索类)资助项目(202103021223029)。
关键词
脑电信号
疲劳检测
栈式自编码
信号特征提取
EEG signal
fatigue detection
stacked auto-encoder
signal feature extraction