Classification of Eye Movement Signals Using Electrooculography in order to Device Controlling

Electrooculography (EOG) signals contain biological information and some existing techniques were developed to analyze these signals for diagnose of illnesses and some of physiological conditions like sleep depth, thought conditions and so on. These signals are usually accompanied by noise and different artifacts. In this paper, some of artifacts are examined to be used as different features of EOG signals. To do this, EOG signals with special and precise physiological experiments, are recorded from 28 healthy people aging from 18 to 26 involving men and women. Then, movement artifacts and other features existing in these signals have been extracted by extreme point strategy and have been classified using K-nearest neighbor algorithm. The goal was controlling some devices such as wheelchair movements to right, left and forward also braking. The proposed method only uses two electrodes and because of its strength in classification, could be a useful method for disabled people.