<![CDATA[Driving safety has become a crucial issue in Indonesia due to the high number of traffic accidents caused by driver fatigue, particularly drowsiness that triggers microsleep and temporary loss of consciousness. This study aims to develop a driver drowsiness detection system based on Arduino Nano and infrared eye blink sensor that detects drowsiness indicators through blink frequency and duration, evaluates the effectiveness of early warnings via buzzer, and identifies technical constraints to enhance reliability in driving simulations. The research method employs a research and development approach with an experimental design, involving 15 participants aged 19-25 years tested in normal conditions and drowsiness simulations for 30 minutes per session, with data collection through eye blink sensor, Arduino Nano, and subjective validation using the Karolinska Sleepiness Scale. Testing shows the system achieves a microsleep detection accuracy of 98% based on eye closure duration exceeding 500 ms and 96% accuracy based on blink frequency below 12 times per minute, with an average response time of 178 ms until buzzer activation, thus capable of providing early warnings before full microsleep occurs. The implications of this research offer a portable, low-cost solution without reliance on complex infrastructure for mitigating accident risks in developing countries, although it still requires optimization against environmental variations such as lighting and head movements for real-world applications on the road.]]>
