DROWTION: Driver drowsiness detection software using MINDWAVE

Abstract

A Significant numbers of road accidents are caused by drowsy driver. This factor can be reduced if the drowsy condition of the driver can be identified and alarmed. This research is conducted by using Electroencephalography approach to detect drowsy state of the driver using Mindwave. Mindwave will sense the value changes of the driver's awareness caused by changes in concentration value. The changes between conscious and drowsiness state are mapped and used as threshold values for triggering the alarm. Result shows that the drowsy state is detected when the average value of low-alpha is below 0.7, the high-alpha value fall below 0.6 and the theta values is below 0.7 from the normal condition. The low-alpha values are sufficient enough to show the condition of drowsiness, but the high-alpha and theta value can be used to minimize the false alarm event. DrowTion application is developed based on this result. DrowTion is implemented with Mindwave headset with the capability of minimizing false alarm and having the capability of giving multiple alarms. Accuracy of DrowTion application in normal condition is about 68,11%.