An Efficient Driver Drowsiness Detection Using Deep Learning

Most of the accidents that happens that happens nationwide is due to the drowsiness. Drowsy driving causes 21% of all fatal collisions. About 168 million adult drivers, or 60% of all drivers, admitted to driving while fatigued in the previous year. Drowsiness is an inattentive state. Drowsiness makes the individual to loss of their control during their driving state, thus leading to devastating accidents. To avoid such catastrophic accidents, a sleep detection system is proposed by which the model embeds an EEG sensor, to detect sleep onset while driving and a CNN-based hybrid model is developed which is used to track the progress of the drivers’ mental and physiological conditions. The activities of brain is evaluated using EEG signal. The EEG signals identifies the fatigue which remains to be the primary reasons for accidents. The model aims to attain the terminology of detecting the driver drowsiness by EEG signals with the high accuracy and efficiency. All the signals are transmitted to the mobile of the driver thereby the CNN algorithm monitors the state of driver. Whenever the driver is in sleepy or fatigue state, then these results are sent to the Car system and the car gives an alert to the driver in the form of an alarm. The combination of CNN based model and EEG sensor will give accurate and optimized results with respect to the driver’s condition. This was the scenario that were often used in the previous works which seems to be a tedious task to wear a hand band and drive for the drivers. So, we are proposing a model that only focuses on the eyes of the driver to detect the onset of the fatigue and drowsiness.