Many crashes with heavy vehicle can be attributed to driver sleepiness or driving impairment due to sleepiness, and it is important to find methods to predict those situations and counteract this problem. The Optalert fatigue management system claims to be able to detect sleepiness. The aims of this study are to (a) evaluate if Optalert can detect sleepiness equally well as other sleepiness indicators and (b) if the data patterns obtained by Optalert correlates with these other sleepiness indicators. Twelve sleep deprived truck drivers drove for about 90 minutes in an advanced moving base truck simulator. The experimental setup, including the sleep deprivation, was designed so that the drivers should become increasingly sleepier during the trial and the intention was that they should fall asleep during the experiment. Four different indicators of sleepiness or driving impairment due to sleepiness were used to monitor the state of the drivers; the Karolinska Sleepiness Scale (KSS), the variability in lateral position (SDLP), the blink duration and the Optalert system. The results show that all four sleepiness indicators increased with time on task. An analysis of variance revealed that the changes were significant for KSS, blink duration and the Optalert system, and a correlation analysis showed that Optalert correlated significantly with blink duration and SDLP. However, even though these correlations were significant, they were all rather low with a maximum correlation coefficient of 0.24. In conclusion, the Optalert system is promising and the sleepiness rating provided by the system works at least equally well as the other three sleepiness indicators. There are some practical limitations to the system; there is no reliable threshold which can be used to determine when a driver is getting too sleepy to drive (this is also the case for other available sleepiness indicators), the driver needs to be attached to the vehicle via the spectacle frames and a wire, and the quality of the eye movement recordings often deteriorated when the driver started driving the truck. Moreover, during the experiment the technical reliability was sometimes low. (Author/publisher) This report may be accessed by Internet users at http://www.vti.se/EPiBrowser/Publikationer%20-%20English/ViP-Publicatio…
Samenvatting