Skip to Main Content
The authors present a study of driver drowsiness, looking for patterns in biomedical and biomechanical variables that allow one to characterise the drowsiness cycle and detect its phases with new technologies. Biomedical signals, eye closure, pressures on the seat, and longitudinal and lateral control of the vehicle were recorded in a driving simulator, during a test in an environment that induced drowsiness, while subjects were motivated to struggle against sleep. Twenty volunteers were measured during the 1 h 45 min tests. A control signal that combined EEG and percent of eye closure (PERCLOS) was defined to classify the different states of the participants during the test. According to that standard, drowsiness was successfully induced in 80 of the subjects. The changes in those states influenced both the performance of the driving task and the biomedical signals, although the former were less sensitive to early fatigue. Heart rate variability and respiration turned out to be promising indicators of the state of the driver, which can be used in future drowsiness detection systems.