ABSTRACT: Driver impairment due to drowsiness is known to be a major contributing factor in many motor vehicle crashes. More than 30% of the road accidents are caused by the fatigue of the driver. At present, there are various drowsiness detection systems available in the market. These systems are implemented using any one of the various implementation techniques such as detection of any behavioral pattern, changes in physiological conditions, or vehicular motion. Consequently, the accuracy of such systems has been found to be low..The paper is built around MCU. Here we are using eye blink sensor. By default the vehicle will be in running condition. During this time if the person closes the eyes automatically the vehicle will be in halt condition, …show more content…
millimeter-wave radar) Limited reaction time Near instantaneous reaction time III. CONCLUSION The system achieved an accuracy above 90 % for all of the scenarios evaluated, including night time operation. In addition, the false alarm rate in the on-the-road area is below 5 %. Our experiments showed that our head pose estimation algorithm is robust to extreme facial deformations.While our system provided encouraging results, we expectthat improving the facial feature detection in challenging situations (e.g., profile faces, faces with glasses with thickframes) will boost the performance of our system. Currently, we are also working on improving the pupil detection using Hough transform-based techniques to further improve the gaze