Detecting Alcohol-Induced Drowsiness in Four Wheeled Vehicles: An Eye Tracking and MQ3 Sensor based Approach with Arduino Uno and Raspberry Pi, Integrated with SMS Notification System and GPS

Alcohol and drowsiness are common reasons for vehicle accidents. For this research, the researchers created an Alcohol and drowsiness detection system with a key ignition access cover and a drowsiness alert system using a speaker to ensure the driver's safety. A notification system was also included to provide additional help to the driver when the subject is either under the influence of alcohol or drowsy during the driving session. A correlation value 0.924 was observed on the MQ3 sensor for maximum values observed on different alcohol concentrations, respectively. A 6-point EAR ratio system was used for the drowsiness detection of the device. It produced an accuracy of 0.7903 and an F-score of 0.8088. GSM Module, the average response time was found to be 4.737 seconds. The average latitude recorded was 14.535411, and the average longitude was 121.000553. The percentage error for longitude is 0.0078, and for latitude is 0.0372, which is close to the actual value of the reference trial.