Electroencephalogram (EEG)-based systems to monitor driver fatigue: a review

• Main Authors: Kamat, Seri Rahayu, Ibrahim, Muhammad Shafiq, Shamsuddin, Syamimi, Md Isa, Mohd Hafzi, Ito, Momoyo
• Format: Article
• Language: English
• Published: Universiti Malaysia Perlis 2022
• Online Access: http://eprints.utem.edu.my/id/eprint/26571/2/0.%20ELECTROENCEPHALOGRAM%20%28EEG%29-BASED%20SYSTEMS%20TO%20MONITOR%20DRIVER%20FATIGUE.PDF; http://eprints.utem.edu.my/id/eprint/26571/; http://dspace.unimap.edu.my/bitstream/handle/123456789/76070/Vol_15_SI_March_2022_365-380.pdf?sequence=1&isAllowed=y

An efficient system that is capable to detect driver fatigue is urgently needed to help avoid road crashes. Recently, there has been an increase of interest in the application of electroencephalogram (EEG) to detect driver fatigue. Feature extraction and signal classification are the most critical steps in the EEG signal analysis. A reliable method for feature extraction is important to obtain robust signal classification. Meanwhile, a robust algorithm for signal classification will accurately classify the feature to a particular class. This paper concisely reviews the pros and cons of the existing techniques for feature extraction and signal classification and its fatigue detection accuracy performance. The integration of combined entropy (feature extraction) with support vector machine (SVM) and random forest (classifier) gives the best fatigue detection accuracy of 98.7% and 97.5% respectively. The outcomes from this study will guide future researchers in choosing a suitable technique for feature extraction and signal classification for EEG data processing and shed light on directions for future research and development of driver fatigue countermeasures.