Determination of Driver's Hypovigilance from Biosignals

Robust and reliable determination of hypovigilance is required in many areas, particularly transportation. Here, new products of Fatigue Monitoring Technologies (FMT) emerge. Their development and assessment requires an independent reference standard of driver’s hypovigilance. Until recently most approaches utilized electrooculography (EOG) and electroencephalography (EEG) combined to descriptive statistics of a few time or spectral domain features, like e.g. power spectral densities (PSD) averaged in four to six spectral bands. Here we present a more general approach of data fusion of many features utilizing computational intelligence methods, like e.g. Support-Vector Machines (SVM). For simplicity, two classes were discriminated: slight and strong hypovigilance. Validation was performed by independent class labels which were assessed from Karolinska Sleepiness Scale (KSS) and from variation of lane deviation (VLD). The first is a measure of subjectively self-experienced hypovigilance, whereas the second is an objective measure of performance decrements. 16 young volunteers participated in overnight experiments in our real car driving simulation lab. Results were compared with PERCLOS (percentage of eye closure), an oculomotoric variable utilized in several FMT systems. We conclude that EEG and EOG biosignals are substantially more suited to assess driver’s hypovigilance than the PERCLOS biosignal. In addition, computational intelligence performed better when objective class labels were used instead of subjective class labels.