EEG-based Drowsiness Detection from Ocular Indices Using Ensemble Classification

One of the reasons for fatal road accidents is sleeping behind the wheel, and there are numerous methods developed to prevent these accidents. The proposed method uses a MATLAB-based BLINKER algorithm to extract ocular characteristics from EEG signals. The classification model makes this method unique to detect drivers' drowsiness using eye blinks. The decision tree model is used in feature selection and a Bayesian optimized ensemble bagged classifier for the prediction. The predictive classification model gives 88.1% accuracy.