Novel Tools for Driving Fatigue Prediction: (1) Dry Eeg Sensor and (2) Eye Tracker

National Sleep Foundation’s Sleep in America (2005) reported 60% of adult drivers driving a vehicle while feeling drowsy in the past year, and more than 37% have actually fallen asleep at the wheel [1]. This paper presented the findings of two novel fatigue prediction tools. The first study presents a 4-channel dry EEG under simulated driving being able to predict when the driver will develop microsleep in the next 10 minutes using only 3 minutes data of collected, with an accuracy of more than 80%. The second study uses an eye tracker to assess the percentage of time that the eyelids were closed (PERCLOS) as a potential marker for fatigue. Results showed that the average magnitude of oscillation (amount of pupil fluctuation), known as Coefficient Magnitude (CM), is generated from real-time wavelet analysis, has the potential to predict fatigue 8-12 minutes ahead with 84% accuracy ahead of compromised driving behavior.