Evaluation of driver fatigue on two channels of EEG data

• Published in: Neuroscience letters Vol. 506; no. 2; pp. 235 - 239
• Main Authors: Li, Wei, He, Qi-chang, Fan, Xiu-min, Fei, Zhi-min
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier Ireland Ltd 11.01.2012
• Subjects: Adult; Automobile Driving; Driver fatigue; EEG; Electroencephalography - methods; Fatigue - complications; GRA; Humans; KPCA; Male; Models, Neurological; Models, Theoretical; Ratio indices; Signal Processing, Computer-Assisted; Young Adult; KPCA; Driver fatigue; Ratio indices; EEG; GRA
• ISSN: 0304-3940; 1872-7972; 1872-7972
• DOI: 10.1016/j.neulet.2011.11.014

► 12 types of energy parameters are computed with 3 bands (α, θ, β) of EEG data. ► A(θ+α)/β,M is confirmed to be optimal indicator of driver fatigue with GRA. ► Fp1 and O1 are determined to be the significant electrodes by KPCA. ► Evaluation model of driver fatigue is established with A(θ+α)/β,M of Fp1 and O1. Electroencephalogram (EEG) data is an effective indicator to evaluate driver fatigue. The 16 channels of EEG data are collected and transformed into three bands (θ, α, and β) in the current paper. First, 12 types of energy parameters are computed based on the EEG data. Then, Grey Relational Analysis (GRA) is introduced to identify the optimal indicator of driver fatigue, after which, the number of significant electrodes is reduced using Kernel Principle Component Analysis (KPCA). Finally, the evaluation model for driver fatigue is established with the regression equation based on the EEG data from two significant electrodes (Fp1 and O1). The experimental results verify that the model is effective in evaluating driver fatigue.