Driving Performance Measures

A review of the previous literature identified driving performance measures that could be used in an experiment to identify an issue of concern. Previous studies have generally used sleep deprivation as the independent measure, and looked at average driving performance across different levels of sleep deprivation (Otmani et al., 2005). Participants attend several sessions in various stage of sleep deprivation, and driving performance is compared across sessions. In these studies, the driver performance was used to compare eye blink behaviour to measure drowsiness or sleepiness. However, it cannot simply be assumed that being sleep-deprived is equivalent to begin drowsy, or that being non-sleep-deprived is equivalent to being non-drowsy (Bittner et al., 2000). Drowsiness changes over time, and can occur when the participants are having a lack of sleep condition as well as normal condition (Suh et al., 2007). One advantage of this research, in comparison to other studies is the measuring of the effect of changing drowsiness over time, rather than simply measuring average driving performance across discrete levels of sleep deprivation.

One problem concerning driving performance measures, as indicators of drowsiness, is inter- and intra- individual differences in driving performance, which could be resolved by a combination of different measures. It has been suggested that the combination of performance measures with physiological measures (e.g. average eye blink durations and blink frequency) would give a sufficiently reliable detection method (Johns, 2003).

4.3.1.

Participants

Eighteen participants were recruited to take part in the simulator test and four of them participated in two sessions, one of which was in a sleep deprived condition. The eighteen participants who attended at least one session comprised fourteen males and four females (age range 20 years to 70 years), with an average age of 35.6 years (SD=12 years). All participants filled out a questionnaire regarding subjective details and sleeping habits before the experiment was run. In addition, four participants who were in a sleep deprived condition (less than three hours during the last 24 hours) were also analysed. The selected participants are very diverse with respect to ethnicity, gender and age. Participants participated in different time sessions and different sleepiness conditions. There were two sessions: session one was in the morning (9-11:45am) and session two was in the afternoon (12-3pm). Each participant drove approximately 30-40 minutes. All the participants had a five to ten minutes training session to get used to the simulator. In order to examine the effects of driver state, the sleep deprived participants participated in the same driving session as the one in which they participated in the good sleep condition. Participants were instructed to drive at a speed of 0 to 60 km per hour, and the simulator speed was set to a maximum of 60kmph.

Several exclusion criteria were applied when recruiting participants. Shift workers and people with diagnosed sleeping disorders were excluded due to the possibility of having disrupted circadian rhythms, and the potential to respond with difficulty after a night of sleep deprivation. Sufferers of severe motion sickness or epilepsy were prevented from participating because of the potential for adverse reactions in the driving simulator. People who wear spectacles to drive also participated and the detection camera was mounted to the spectacles. People with narrow eyes were also included because their eye-sclera shape helped to calibrate the detection system. All the sleep deprived participants were volunteered are instructed not to have a morning sleep. All participants were advised not to take any drugs or alcohol before the simulation test. Finally, transport was arranged for the participants who volunteered in a sleep deprived condition.

4.3.2.

Driving Scenarios

The simulation test consisted of a two-lane rural road, with centre lines and lateral edge lines. All participants were asked to take part in a driving simulator-based experimental trial in which they were required to drive along a scenario partitioned into a “rural” section, representing dual lane country roads with hills and bends, a “straight” section, representing a stretch of same road, fairly straight, open road, free of traffic. The participants were seated in the driving simulator at a distance of 1.5 metres from the screen. The simulation is intentionally designed to increase the sleepiness of the participants by inclusion of the above effects. The road edges and out-of-bounds were created as rough surfaces, which provided an audio feedback to participants (participants) if they ran off the road. It was considered that having the rough road edges was a more realistic option than having the experimenter verbally

advise the participant that they must return to the road. The driving scenario was a loop of approximately two to three minutes to continue one loop, with average speed of 60 km/h. Participants were free to drive at any speed between 0 to 60 km/h. An on-screen timer (clock) was available. The signed speed limit for most of the scenario was 60 km/h because at greater speeds it is difficult to control the vehicle on the centre line of the road. The single driving session for each participant tested lasted approximately 30-40 minutes.