IDENTIFICATION OF DRIVER'S DROWSINESS USING DRIVING INFORMATION AND EEG

Abstract: The paper summarizes the flrst results of an identiflcation of sleepystate of drivers using a complex set of outputs from simulated driving. The drivinginformation, such as deviation from the centerline of the road and the steeringwheel position as well as two-point EEG, was used. The process consists of thepreprocessing of data, in fact a transformation into a form proper for classiflcation,and a classiflcation into one of two classes, i.e. wakefulness and drowsiness. Therewere two groups of drivers submitted to tests, the wakeful ones, and the driversafter serious sleep deprivation. We found that it is possible to distinguish thesegroups using an appropriate classifler with some rather substantial error, whichcan possibly be tackled by using an apt methodology.Key words: Wakefulness, drowsiness, sleepy state of drivers, classiflcation,driving simulationReceived: September 6, 2010Revised and accepted: December 13, 2010 1. Introduction Driver’s drowsiness is a very dangerous phenomenon causing approximately sevenper cent of all car accidents. But it is impossible to give an absolutely precise num-ber of these cases because reasons for such road accidents may also be classifledas \inattentiveness to primary driving tasks". Drowsiness is analyzed or identifledby a lot of approaches based on EEG analysis ([1]) and other psycho-physiologicalmeasures, often also on analysis of driver’s responses while controlling the car. Re-cently there has also been a noticeable progress in face analysis ([2]). However, forpractical applications the main problem lies in distinguishing between the driver’swakeful and drowsy state in the car or truck.