Enhancing driver drowsiness detection for data acquisition stage using electrocardiogram

• Main Authors: Nor Shahrudin, Nur Shahirah, Sidek, Khairul Azami, Nazmi Asna, Nur Aaina Nazihah, Nordin, Anis Nurashikin, Jalaludin, Muhammad Rasydan
• Format: Conference or Workshop Item
• Language: English; English; English
• Published: IEEE 2021
• Subjects: TK7885 Computer engineering
• Online Access: http://irep.iium.edu.my/90588/8/90588_Enhancing%20driver%20drowsiness%20detection%20for%20data%20acquisition%20stage%20using%20electrocardiogram.pdf; http://irep.iium.edu.my/90588/7/90588_Programme%20Schedule.pdf; http://irep.iium.edu.my/90588/19/90588_Enhancing%20driver%20drowsiness%20detection_Scopus.pdf; http://irep.iium.edu.my/90588/; https://ieeexplore.ieee.org/document/9467209

Road accidents can occur based on many factors and one of them is due to driver drowsiness. These fatalities could cause death which affects our country’s economy. Thus, this study proposed a driver drowsiness detection based on Electrocardiogram (ECG) for the data acquisition stage. ECG has been used in collecting data from the human body that used electrodes and place it on human skin to detect the electrical activity of the heart. This study proposed a drowsiness detection through ECG signal involving 10 subjects aged in their early 20s regardless of their gender. All subject used for this test is free from any kind of drugs, alcohol or even caffeine. The ECG data were collected from a source called The ULG Multimodality Drowsiness Database (DROZY). Next, the signal obtains from the database does not need to undergo the filtering process since the R-peak of the data can easily be detected. The feature that has been extracted is the R peak so the HRV analysis can be used to classify the state of the subject, either awake or drowsy. Other than that, the data of the cardioid of each subject also being measured and the Euclidean distance of it being compared. The outcome of this study shows that the amplitude of the drowsy phase will be lower compared to the normal state and the same goes for the Euclidean distance of Cardioid based graph.