Driver Drowsiness Detection Using Smartphone Application

Abstract

The risk of road accidents increases due to tiredness resulting from long-distance driving and sleep deprivation, leading to tiredness in the driver. To address this issue, a proposed framework suggests a smart phone based system that uses a three-stage approach for detecting drowsiness. In the first stage, the front camera captures images and uses a modified eye state classification method to measure the percentage of eyelid closure (PERCLOS), which is supplemented with near-infrared lighting for night driving. In the second stage, the microphone records speech data to determine the voiced to unvoiced ratio if PERCLOS crosses a threshold. In the third stage, the driver is required to touch the screen within a certain time to confirm their alertness, triggering an alarm if deemed drowsy. The device also maintains a file of the metrics and coordinates. When compared to the existing systems, the proposed method has three advantages: a more reliable three-stage verification process, implementation on readily available Android smartphones, and SMS alerts to the control room.