Development of low-cost embedded-based electrooculogram blink pulse classifier for drowsiness detection system

This paper discusses the development of a low-cost embedded-based electrooculogram (EOG) blink pulse classifier. A signal conditioning circuit from a single quad operational amplifier (Op-Amp) and an Arduino based on the ATmega32u4 AVR 8-bit microcontroller board comprised the major components of the embedded-based classifier. The evaluation of the nearest neighbor algorithm classifier resulted to an accuracy of 87.14%, precision of 93.33% and sensitivity of 80.00%. Further, based on the participants who evaluated the drowsiness detection system the results were 3.38 and 4.13 with verbal interpretations of comfortable and very convenient respectively.