Estimating prediction horizon of driver fatigue using Euclidean distance-based similarity score between electroencephalograms

• Main Authors: Rodney Petrus Balandong, Syaimaa Solehah Mohd Radzi, Zulkifli Yunus, Mohamad Zul Hilmey Makmud, Tang, Tong Boon
• Format: Article
• Language: English
• Published: e-VIBS, Faculty of Science and Natural Resources 2024
• Subjects: Q1-390 Science (General); RC346-429 Neurology. Diseases of the nervous system Including speech disorders
• Online Access: https://eprints.ums.edu.my/id/eprint/43023/1/FULL%20TEXT.pdf; https://eprints.ums.edu.my/id/eprint/43023/

Driver fatigue is one of the major causes of road accidents. While numerous electroencephalography (EEG) related methodologies have been proposed for automatic fatigue detection, very little attention has been given to explore the use of EEG in the estimation of the prediction horizon of driver fatigue. This paper proposed a novel framework based on the similarity score measured by the Euclidean distance in the brain oscillatory rhythmic patterns to determine how far ahead the decrement in driver’s vigilance could be detected. A new metric for the confidence level of the estimation was also suggested to quantify prediction reliability. The proposed framework was assessed using the data from a driving simulation experiment involving 20 healthy female subjects with mean age of 22 and found that the prediction horizon can be extended up to 56s solely based on EEG features. In conclusion, this study demonstrated how the EEG features can be used for the estimation of prediction horizon in driver fatigue management.