Driving Fatigue Classified Analysis Based on ECG Signal
The ECG data obtained through experiment is divided into normal state and fatigue state two types by obtaining ECG signal under different conditions of human through experiments and selecting PERCLOS value as basis to judge the degree of fatigue under controlled environment. On the basis, use Kernel Principal Component method to investigate the selected ECG signal parameters whether can effectively express the state of human fatigue. Analyzing the collected samples by using Kernel Principal Component method shows that selecting appropriate kernel function and related parameters can effectively separated normal samples and fatigue samples and that it is feasible to detect fatigue through the selected ECG signal parameters. Meanwhile, fatigue divisibility of ECG signal linear parameters was similarly analyzed without considering nonlinear parameters, the results show that only using the linear parameters could also monitor the degree of fatigue, but the boundary of samples is not much more obvious than the boundary of integrated linear and nonlinear information.
ECG kernel principal component SVM driving fatigue
Qun Wu Yangyang Zhao Xiangang Bi
School of Art and Design Zhejiang Sci-Tech University Hangzhou, China
国际会议
杭州
英文
1122-1125
2012-10-28（万方平台首次上网日期，不代表论文的发表时间）