Mobile EEG Based Drowsiness Detection using K-Nearest Neighbor

Abstract

In this research, a drowsiness detection system, named Drowsiver, was developed for a mobile electroencephalograph (EEG) and a mobile phone. The system is expected to minimize the causes of accidents caused by drowsy drivers. By using Electroencephalogram (EEG), the condition of drowsiness is detected by recording the electrical activity that occurs in the human brain and is represented as a frequency signal. The signal is transmitted to the Android mobile application via Bluetooth and will give an alarm notification if the drowsiness is detected. The brainwave from the mobile EEG is processed using Fast Fourier Transform (FFT) to extract its features. These features are classified using K-Nearest Neighbor (KNN) classifier. The system produces the best performance with the highest accuracy of 95.24% using the value of k=3 and four brain waves as features, namely Delta, Theta, Alpha, and Beta waves.