【摘要】 针对在驾驶疲劳状态识别研究中采集的脑电信号(Electroencephalograph,EEG)具有非线性特点,以及目前基于EEG的驾驶疲劳状态识别方法仍存在识别准确率不很理想的问题,提出基于样本熵和核主成分分析的驾驶疲劳状态识别方法。所研究的方法结合了基于样本熵在驾驶疲劳状态识别方法具有较好的识别准确率优势以及核主成分对主成分是非线性的情况下具有较好的降维效果优势。实验结果表明所提出的方法是有效的。更多还原

【Abstract】 The EEG data acquired in the study of driving fatigue state recognition has nonlinear characteristics,and the current recognition method based on EEG for driver fatigue state still has the problem that the recognition accuracy rate is not ideal.This paper proposes a driving fatigue identification method based on the sample entropy(SE) and the kernel principal component analysis(KPCA).The proposed method combines the advantages better recognition accuracy of SE in the recognition of driving fatigue state and the benefits of satisfactory dimensionality reduction of KPCA.The experimental results show that the proposed method is effective.更多还原