Blink waveform in electrooculogram (EOG) data was used to develop and adjust the method of drowsiness detection in drivers. The origins of some other waveforms in EOG signal were not very clearly understood. The purpose of this thesis work is to study the EOG signal and give explanation of different kind of waveforms in EOG signal, and give suggestions to improve the blink detection algorithm.

The road driving test video records and synchronized EOG signal were used to build an EOG library. By comparing the video record of the driver’s face and the EOG data, the origin of the unknown waveforms were discovered and related with the driver’s behavior. Literature descriptions were given to explain the EOG signal. The EOG library is the main result of this project. It organized by different types of EOG signal. Description and explanation were given for each type of waveform, as well as some examples. The knowledge gained from the previous research review and the EOG library gives some improvement suggestions for the blink detection algorithm. These suggestions still need to be verified in practical way.