Driving Drowsiness Detection Using fusion of EEG, EOG and Driving Quality Signals

This study investigates the detection of the drowsiness state for a future application such as in the reduction ofthe road traffic accidents. The Electroencephalography(EEG), Electrooculography (EOG), Driving Quality (DQ), and Karolinska Sleepiness Scale (KSS) data of 7 male during approximately 20 hours of sleep deprivation were recorded. To reduce the eye blink artifact, an automatic mechanism based on the Independent Component Analysis (ICA) method and Higuchi’s fractal dimension has been applied. Afterrecordings, for selecting the best subset of features, a new combined method, called Class Separability Feature Selection- Sequential Feature Selection (CSFS-SFS), has been developed. This method reduces the time of calculations from 6807 to 2096 seconds(by 69.21%)while the classification accuracyremain relatively unchanged. For diagnosis of the drowsiness state and classification of the state, a new approach based on a Self Organized Map (SOM) network is used. First, using the data obtained from two classes of awareness state (AS) and drowsiness state (DS), the network achieved an accuracy of 76.51±3.43%. Using data from three classes of AS, AS/DS (passing from awareness to drowsiness), and DS to the network an accuracy of62.70±3.65% was achieved. It is suggested that the drowsiness state during driving is detectable with an unsupervised network.