Abstract
Driver fatigue is a major cause of road accidents and automatic fatigue detection can help preventing these injuries. Electroencephalogram (EEG) microstate analysis has gained popularity as a tool for detecting brain state, mental workload and brain disease. The aim of this research is analyzing the EEG microstate features to effectively detect the driver fatigue state based on microstate features and Support Vector Machine (SVM) classifier. The global field power of EEG and its local maximum are calculated and then clustered in to four microstates. Four features were calculated for each segment of the data including duration, occurrence, time coverage and power. The extracted features in conjunction with SVM classifier have been used for automatic detection of fatigue state. The quantitative results based on leave-one-out approach using EEG data of 10 healthy subjects show that the proposed method has accuracy of 75%. To examine the optimal region of the brain and electrode selection, we divided the electrodes into four distinct regions and evaluated the accuracy of fatigue detection for each region. Our findings indicate that the central region yielded the best results.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call