Development of Detection System Using EEG and ECG To Determine an Onset of Driver Drowsiness Due to Sleep Deprivation

Drowsy driving is widely recognized as a prominent contributor to fatal vehicular incidents. Regarding these issues, several measuring and detection techniques rely on the physiological motion of the driver. The present study investigates the identification of fatigued drivers resulting from insufficient sleep by the analysis of electroencephalogram (EEG) and electrocardiogram (ECG) signals. The data collection will involve a cohort of 10 individuals, all of whom are in good health and fall between the age range of 22 to 27. The study involved the completion of five trials over the course of four separate days, during which participants were required to operate a driving simulator in both an alert and drowsy condition. The data underwent statistical analysis, and signal pre-processing techniques will be utilized to build suitable data validation procedures. The level of drowsiness was assessed by calculating the average of the frequency band for EEG signals, the average of the RR interval, and the LF/HF ratio for ECG signals obtained from the participant’s data. The classification performance of the SVM and kNN classifiers yields an accuracy of 77.78% for SVM and 80% for kNN in the case of EEG data. Similarly, for ECG data, both SVM and kNN classifiers achieve an accuracy of 83.33% when using the same set of features. The introduction of multimodal signals can enhance the outcomes of categorization in response to the scarcity of subjects. The utilization of multimodal signals is expected to yield superior outcomes. The implementation of a sleepiness detection system utilizing the Internet of Things (IoT) capabilities and an embedded-based platform is undertaken. The successful development of the sleepiness detection system resulted in the creation of a wearable gadget that had the capability to notify and alert the user.