A Hybrid Approach to Detect Driver Drowsiness Utilizing Physiological Signals to Improve System Performance and Wearability
Centre for Intelligent Signal and Imaging Research, Universiti Teknologi PETRONAS (UTP), Seri Iskandar 32610, Malaysia
Department of Electrical, Electronic, and Information Engineering Guglielmo Marconi, University of Bologna, Bologna 40126, Italy
Department of Electrical and Electronics Engineering, Universiti Teknologi PETRONAS (UTP), Seri Iskandar 32610, Malaysia
Author to whom correspondence should be addressed.
Sensors 2017, 17(9), 1991; https://doi.org/10.3390/s17091991
Received: 23 April 2017 / Revised: 29 June 2017 / Accepted: 10 July 2017 / Published: 31 August 2017
(This article belongs to the Special Issue Sensors for Transportation)
Driver drowsiness is a major cause of fatal accidents, injury, and property damage, and has become an area of substantial research attention in recent years. The present study proposes a method to detect drowsiness in drivers which integrates features of electrocardiography (ECG) and electroencephalography (EEG) to improve detection performance. The study measures differences between the alert and drowsy states from physiological data collected from 22 healthy subjects in a driving simulator-based study. A monotonous driving environment is used to induce drowsiness in the participants. Various time and frequency domain feature were extracted from EEG including time domain statistical descriptors, complexity measures and power spectral measures. Features extracted from the ECG signal included heart rate (HR) and heart rate variability (HRV), including low frequency (LF), high frequency (HF) and LF/HF ratio. Furthermore, subjective sleepiness scale is also assessed to study its relationship with drowsiness. We used paired t-tests to select only statistically significant features (p < 0.05), that can differentiate between the alert and drowsy states effectively. Significant features of both modalities (EEG and ECG) are then combined to investigate the improvement in performance using support vector machine (SVM) classifier. The other main contribution of this paper is the study on channel reduction and its impact to the performance of detection. The proposed method demonstrated that combining EEG and ECG has improved the system’s performance in discriminating between alert and drowsy states, instead of using them alone. Our channel reduction analysis revealed that an acceptable level of accuracy (80%) could be achieved by combining just two electrodes (one EEG and one ECG), indicating the feasibility of a system with improved wearability compared with existing systems involving many electrodes. Overall, our results demonstrate that the proposed method can be a viable solution for a practical driver drowsiness system that is both accurate and comfortable to wear.