Real-time EEG-based mental fatigue estimation model

Numerous high-risk professions require sustained states of alertness. It is estimated that nearly one-third of traffic accidents worldwide may result from inattention and fatigue while driving. Developing methods to prevent as many accidents as possible caused by inattention and fatigue is essential. This conference paper focuses on real-time fatigue estimation by measuring electrical brain activity. A non-invasive method of recording brain activity, using a device called an electroencephalograph, was employed. The study of mental fatigue and its estimation is an area of interest for many studies for several years. Electroencephalography allows the estimation of mental fatigue in a short-term signal, making it an ideal method for this study. We focused on the beta/(theta+alpha) ratio, which is identified as the fatigue index. A simple yet effective model was developed in Matlab Simulink, using power spectral density method to calculate the fatigue index. Measurements were conducted in a way to simulate monotonous driving in a car on a highway. The model can estimate mental fatigue level within the first ten seconds of real-time measurement. For future studies, the focus should be on the technical design and development of a wearable single-electrode electroencephalograph. This device may be used for the estimation of mental fatigue levels in real-life situations, making it suitable for high-risk professions and everyday driving.