To investigate the correlation between electroencephalogram (EEG) and driving fatigue states, this study used machine learning algorithms to detect driving fatigue based on EEG. 14 channels of EEG data were collected from thirty-four healthy subjects in this research at Northeastern University. Each subject participated in two scenarios (baseline and fatigue scenarios). Subjective ratings of fatigue levels were also obtained from the subjects using the NASA-Task Load Index (TLX). The common spatial pattern (CSP) algorithm was used to extract features from the raw EEG data. The support vector machine (SVM) was used as the classifier in the design of the machine learning algorithm. A grid search cross validation was exploited to find optimal hyperparameter settings. The best classification result was 90%, obtained by using all 14 EEG channels and linear kernel of SVM. The experimental results proved that a machine learning algorithm was able to reliably classify driving fatigue states using EEG data. This study demonstrated that CSP and SVM were promising in detecting driving fatigue, and therefore, they could be strong foundations for future efforts to reduce traffic accidents and save thousands of human lives.
WANG, LI; JOHNSON, DAVID; and LIN, YINGZI
"Using EEG to detect driving fatigue based on common spatial pattern andsupport vector machine,"
Turkish Journal of Electrical Engineering and Computer Sciences: Vol. 29:
3, Article 7.
Available at: https://journals.tubitak.gov.tr/elektrik/vol29/iss3/7