Automatic removal of ocular artifacts in EEG signals for driver’s drowsiness detection: A survey

Abstract

Nowadays, there are many causes of daily traffic accidents, one of them is the loss of concentration while driving due to drowsiness, followed by sleep while driving. Sleeping means the nap; no more than a few seconds, but it is enough to create a traffic accident, in which the driver and the clashes may lose their lives. In this context, there is a host and considerable research efforts made in designing driver monitoring systems with the aim to reduce the vehicular accidents posing a challenging issue for the society. Several drowsiness detection techniques have been proposed in the literature, including artificial neural network, image processing, and physiological measurement techniques. Among the proposed solutions, the electroencephalographic (EEG) measurement is one of the reliable techniques. Nevertheless, the brain signals can be easily contaminated by many artifact types arising from cardiac (ECG), muscles (EMG), and ocular activities (EOG). From these physiological artifacts, ocular activities are one of the most eminent over other noise sources. In this paper, we provide a comprehensive survey, which covers the vast existing techniques of removing ocular artifacts from EEG recordings for driver's drowsiness detection systems. We propose a novel taxonomy of solutions and we compare them with respect to relevant criteria.