To be or not to be… awake? A comparison of subjective and objective methods for drowsiness detection in drivers

Drowsiness or fatigue poses a significant risk for road accidents and carries important consequences for overall road safety. Timely warnings to drowsy drivers can be crucial to prevent potentially fatal accidents. By now, the techniques employed for assessing drowsiness can be subjective, such as questionnaires, or objective, such as the analysis of the behavioral conditions of the driver or the vehicle and the physiological parameters of the driver. As a matter of fact, the most robust method for detecting drowsiness is an objective one, in particular the analysis of physiological parameters through full polysomnography. However, the best practice for road safety still utilizes subjective parameters to establish the drowsiness of the driver. A major question still exists whether subjective sleepiness can predict actual fall-asleep events while driving (e.g., micro-sleep events). In this study, a comprehensive comparison between objective and subjective methods for drowsiness assessment is provided by the use of medical-grade devices and a high-fidelity driver simulator design. As a result, the unpredictability and unreliability of subjective methods for drowsiness assessment were demonstrated.