Wristband-Type Driver Vigilance Monitoring System Using Smartwatch

Boon Giin Lee, Boon Leng Lee, Wan Young Chung

Research output: Journal PublicationArticlepeer-review

38 Citations (Scopus)

Abstract

Studies have presented that the driver vigilance level has serious implication in the causation of road accidents. This paper focuses on integrating both the vehicle-based control behavior and physiological state to predict the driver vigilance index which is evaluated by using a smartwatch. The vehicle control behavior can be observed from the steering wheel movement. Our study utilized the smartwatch motion sensors to study the steering wheel behavior. Meanwhile, physiological state of driver reflects the driver capability of safety alert driving which is estimated by photoplethysmogram (PPG) and respiration signals in this paper. The PPG sensor is integrated in a sport wristband with a Bluetooth low energy module, transmitted the PPG signals to smartwatch in real time. The steering angle is derived by the reading from smartwatch built-in accelerometer and gyroscope sensors. On the other hand, the respiration is derived using the PPG peak baseline method. In order to utterly investigate the sleepiness-induced factors, the time, spectral, and phase space domain features are calculated. Considering the smartwatch processing capability, mutual-information technique is applied to designate the ten most descriptive features. Then, the extracted descriptive features are serve as parameters to a classifier to determine the driver aptitude status. The features are analyzed for their correlation with the subjective Koralinska sleepiness scale and through recorded video observations. The experimental results reveal that our system is capable of estimating driver hypervigilance at average of 96.5% accuracy rate by evaluating on both driving behavior and driver physiological state, provided a novel and low-cost implementation.

Original languageEnglish
Article number2447012
Pages (from-to)5624-5633
Number of pages10
JournalIEEE Sensors Journal
Volume15
Issue number10
DOIs
Publication statusPublished - 1 Oct 2015
Externally publishedYes

Keywords

  • Bluetooth
  • accelerometer sensor
  • gyroscope sensor
  • motion sensors
  • photoplethysmogram
  • smartwatch
  • wristband

ASJC Scopus subject areas

  • Instrumentation
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Wristband-Type Driver Vigilance Monitoring System Using Smartwatch'. Together they form a unique fingerprint.

Cite this