Fractal dimension based electroencephalogram analysis of drowsiness patterns

Abstract

As drowsiness is one of the prime causes of traffic accidents, monitoring drivers' drowsiness is an active safety-focused research which involves monitoring both physical and physiological changes. This paper aims to characterize a subject's drowsiness based on electroencephalogram (EEG) analysis. The two effective fractal dimension (FD) algorithms: the variance fractal dimension (VFD) and the detrended fluctuation analysis (DFA) were investigated to reveal these EEG patterns. EEG data were recorded from sixteen channels of four healthy male subjects aged 19-33 years. Our result demonstrated that the proposed algorithms feasibly recognized alertness and drowsiness of EEG waveforms.