Real Time Tracking of Driver Fatigue and Inebriation Maintaining a Strict Driving Schedule

This paper is concerned about the methods of road safety by addressing potential causes such as drowsiness and inebriation maintaining a strict schedule by recognizing the driver's face. Increasing unawareness towards traffic rules yields more and more accidents by the day. Drowsiness results from the monotony towards driving and inebriation results from the unawareness or unwillingness to abide by the traffic rules. This conundrum victimizes both the person inside and outside the vehicle. However, drowsiness prevention requires a method of detecting the deterioration of the vehicle operator's attention in a legitimate way along with an alerting mechanism. Though the existing solutions are developed through some unique methods, there are still some issues addressing yawn, blink issues, and alcoholism which have not been considered in their systems. This study aims to develop an improved and innovative approach to solving this issue. A train model developed by histogram oriented gradient (HOG) and linear support vector machine (SVM) extracts the eye and mouth position and calculates the eye aspect ratio (EAR), mouth aspect ratio (MAR) and MQ-3 sensor for measuring the degree of concentration of alcohol in the air. These data are then compared with the threshold value which is developed from a data-set of the aspect ratio of sleeping or drowsy face models.