A Performance Study of Drivers Through the Development of a Microsleep Prevention System

Most of the detection systems in use so far, for example, eye tracking cameras and EEG, are usually costly, difficult for general users. Therefore, low- cost and non-intrusive solutions that can be easily integrated towards the users of vehicle is greatly needed. For this reason, this paper seeks to develop a practical microsleep detection system using sensors attached to a steering wheel that drivers can afford. The pressure from holding the force controller is detected with an FSR and steering is sensed by the MPU6050 accelerometer. Both sensors are linked to an Arduino Uno microcontroller and attached to the head using Velcro and foam padding to prevent making any permanent changes. The program is designed to create an alert with an LED and buzzer when weak grip occurs at the same time as steering, suggesting that microsleep might be about to happen. Also, the prototype is tested over and over in simulations to check both the sensors’ reliability and how quickly their responses are. Based on observations, the initial tests results show the characteristic normal driving and when the users start losing focus. On average, the frequency of loose grip increases significantly after 15 min. As a result, this research was able to observe fatigue behaviour with participant 2 showing the highest fatigue level, 1012 instances of weak grip and 568 cases of erratic steering. Overall, this research manages to outline the effectiveness of the detection algorithm.