By: Thomas X.
School: Yorba Linda High
Science Teacher: Daniel Eliot
Drowsy driving can be lethal. Although, determining a precise number of drowsy-driving crashes, injuries, and fatalities is not yet possible, the National Highway Traffic Safety Administration (NHTSA) estimates that in 2017, 91,000 police-reported crashes involved drowsy drivers. These crashes led to an estimated 50,000 people injured and nearly 800 deaths. While drowsy driving and accidents related to drowsy driving inevitable, they may be preventable.
This study aims to address this problem through a mobile application that uses artificial intelligence to detect drowsiness in the driver. The app is able to detect the eyes of the user once he or she starts the app, captured by the phone’s camera. The algorithm would then make calculations based on the user’s eye openness and closed frequencies to determine the drowsiness. When the drowsiness data reaches a threshold, the algorithm will trigger an alarm from the mobile phone.
For this experiment, the app was applied to actual driving environments -both day and night- on the streets, and a normal control situation and conducted a qualitative evaluation approach. The results shows that it is 90% effective during day time and 85% effective during night time. While effectiveness and accuracy for night time detection and detections with the presence of other distractions can be further improved, these results show promise for both the app and the prevention of catastrophic losses of human life.