Automated vigilance classification based on EOG signals: Preliminary results

Abstract

The dramatic increase in the number of traffic accidents associated with sleepiness has stimulated research aimed at identifying behavioural and physiological correlates of alert ness in different occupational settings ([1]). The paper describes the first results of a study working on a model for a automated vigilance classification. First measurements of EOG and EEG signals have been performed on persons doing the Mackworth Clock test. In this paper the analysis of the EOG parameters is described and the results concerning the correlation with the re action time of the Clock test are shown and discussed. The EOG is so far the preferred biosignal because it is relatively easy to measure without to much burden on the user. It is thinkable that a smart device based on EOG recording will serve as an all solution for warning the user in the case of micro sleep events or extreme vigilance decline. In future it would be very important to have models and algorithms available giving the possibility of a automated real time vigilance classification based on biosignals. A device available on the market giving a feedback to the driver when the vigilance is decreasing would contribute to in crease the safety in car traffic. © 2009 Springer-Verlag.