Smartwatch-Based Wearable and Usable System for Driver Drowsiness Detection

Abstract

Drowsiness is one of the leading causes of near-miss or real road accidents. Researchers have invested a considerable amount of effort identifying ways to the detect drowsiness state of drivers and alert them in a timely manner to avoid serious consequences. Recent works on drowsiness detection have been focused on proposing wearable solutions that can be portable and used by the driver with ease. Unfortunately, majority of these are difficult to use on a daily basis. In this paper, we propose a usable and wearable solution that tracks the user’s state of activeness using a smartwatch and gives them real-time feedback. Our proposed solution measures the Heart Rate Variability (HRV) coupled with Galvanic Skin Response (GSR) to detect whether the driver is drowsy behind the wheel or not. HRV measures the fluctuations between the heart beats whereas GSR measures the emotional arousal from skin’s sweat gland activity. An auditory feedback is provided to the driver if the HRV and GSR values are found below the expected thresholds. The system demonstrated an accuracy of 80%, precision of 97% and recall of 82%. Furthermore, we also conducted a usability study to assess the acceptance of the proposed application.