Drowsy drivers have been identified as the main leading cause of car accidents. It is estimated that there are 76,000–100,000 car crashes occurring each year in the United States, leading to 1500 deaths and thousands of injuries (Knipling and Wang, 1995; Wang et al., 1996). Drowsy drivers cannot focus on driving and tend to commit on manipulating errors.
Their information processing speed and working memory capacities are decreased and drastic changes on their task performance occurs (Wylie et al., 1996; Chang and Mannering, 1999;
Kostyniuk et al., 2002; Hendrix, 2002). Through face to face interviews with 593 long-distance drivers, McCartt reported that 47 % of the respondents had ever fallen asleep and 25.4% had fallen asleep during driving of the past year (McCartt et al., 2000). Several factors contribute to the occurrence of symptoms of fatigue and falling asleep in drivers, such as lack of sleep, long driving hours, driving in a monotonous environment, taking sedative drugs or drinking alcohol before driving and driving at midnight, early morning, or mid-afternoon hours. Therefore, accurate and non-intrusive real-time monitoring of driver's drowsiness would be highly desirable, particularly if this measurement could be further used to predict changes in driver's performance capacity.
1.1. Current researches of drowsiness
There are several ways to detect drivers’ drowsiness. For example, it can be directly captured from video images (Summala et al., 1999), the rate and duration of the EOG (electrooculogram, Horne and Reyner, 1996). It can also be estimated from bio-signals such as ECG (electrocardiogram), body pressure, and respiration (Milosevic, 1997; Chung et al., 1999), and the electroencephalogram (Horne and Reyner, 1995; Khardi and Vallet, 1994; Lal and Craig, 2002, Huang et al., 1996; Vuckovic et al., 2002; Roberts et al., 2000; Khalifa et al.,
The abundant information in EEG recording can be related to drowsiness, arousal, sleep, and attention (Santamaria and Chiappa, 1987). Previous studies showed that changes in the EEG theta band and the alpha band reflect cognitive and memory performance (Klimesch, 1999). For example, Makeig and Jung (1996) and Huang et al. (2005) reported that mean activity levels in the (< 4 Hz) delta and (4-6 Hz) theta bands, and at the sleep spindle frequency (14 Hz) as well as the baseline alpha band power were significantly increased from alert to poor/drowsy performance. Several EEG studies related to driving also suggested that alpha-band and theta-band power increased as the alertness level of the driver decreased (Torsvall and Akerstedt, 1987; Eoh et al., 2005; Otmani et al., 2005). Though many studies on the driver’s drowsiness with EEG have been performed, the driving simulation apparatus of experiments in the literatures are mostly constructed only on the monitors. But, the static driving simulation is difficult to approach the realistic driving condition, such as the vibrations that would be experienced when driving an actual vehicle on the road.
1.2. Kinesthetic perception during driving
The driving motion is one of the most experienced kinesthetic perceptions in our life, in other word, the perception we sensed during the vehicle speed or direction change. Whenever the vehicle accelerates, decelerates or curves in a corner, we experience a force pulling our body against the direction of moving. For a driver, the perception to motion includes kinesthetic and visual stimulus. A driver does not sense only the pushing or pulling his/her body by a force, but also the scene change related to vehicle movement. The driving perception includes the co-stimulation of visual cue, vestibular stimulation, muscle reaction and skin pressure. It is indeed a complicated mechanism to understand.
There are numbers of difficulties in investigating the driving perception. First of all, the safety of subject must be guaranteed. Experiments should be held under a safe driving
environment, it is very dangerous to conduct driving experiments on the road. Second, appropriate monitoring and data acquisition are needed to study the influence of kinesthetic stimuli. The stimulation should be simple enough and repeatable to keep experiment under control. Third, objective evaluation should be assessed in the studies.
One of the solutions is to conduct driving experiments using a realistic simulator, which is widely used in driving related researches (Kemeny and Panerai, 2003). For the necessity of motion during driving, literatures showed that the absence of motion information increased reaction times to external movement perturbations (Wierville et al., 1983), and decreased safety margins in the control of lateral acceleration in curve driving (Reymond et al., 2001).
In real driving, improper signals from disordered vestibular organs were reported to determine inappropriate steering adjustment (Page and Gresty, 1985). Moreover, the presence of vestibular information in driving simulators shows the importance for it influences the perception of illusory self-tilt and illusory self-motion (Groen et al., 1999). All the above studies emphasized the importance of motion perception during driving with the assessment of driving performance and behavior. Our previous studies also demonstrated that multiple cortical EEG sources responded to driving events differentially in dynamic and static environment. Specifically, the alpha band variations occurred in many components (Mu, parietal and occipital) during driving, especially when the vehicle is moving. It is still unclear to what extent the kinesthetic stimulation would interfere with the fluctuations of driver's global level of drowsiness accompanying changes in driver's performance.
1.3. Virtual reality dynamic simulator
Virtual reality (VR) technology is gradually being recognized as a useful tool for the study and assessment of normal and abnormal brain function, as well as for cognitive rehabilitation. Virtual Environments (VE) are created by powerful computers that generate
realistic animated graphics in three dimensions. Creating carefully controlled, dynamic, 3D stimulus environments combined with physiological and behavioral response recording can be offer more assessment options that are not available by traditional neuropsychological methods.
The VR technique allows subjects to interact directly with a virtual environment rather than monotonic auditory and visual stimuli. It is an excellent strategy for brain research on interactive and realistic tasks due to low cost and avoiding risk of operating on the actual machines. In recent years, some researchers designed the VR senses to provide the appropriate environments for brain activity study (Bayliss and Ballard, 2000; Eoh et al., 2005;
Huang T.Y et al., 2005). Integrating the VR scene with dynamic motion platform is excellent for studying the influence of kinesthetic stimulus on cognitive state. Therefore, a VR-based dynamic motion platform combined with EEG measured system is an innovation in brain and cognitive engineering researches.
1.4. Aims of this thesis
Aims of this thesis were (1) to characterize EEG changes with the degradation of the alertness and (2) to assess EEG dynamics in responses to kinesthetic stimulus in different cognitive states. We first constructed a Virtual-Reality interactive driving environment consisting of a highway scene and a six degree-of-freedom (6-DOF) motion platform. Then, we designed a lane-keeping driving experiment to indirectly quantify driver’s drowsiness level (Philip et al., 2003). Therefore, we could easily demonstrate that changes of EEG activities were correlated with driver’s response performance as well as the influences of kinesthetic stimulation on EEG dynamics from alter to drowsiness. Accordingly, this thesis provided strong evidences to show that the dynamic motion platform is required for correctly estimating driver’s cognitive states under driving in the future.
2. Materials and Methods