Skip to main content
placeholder image

Automatic driver cognitive fatigue detection based on upper body posture variations

Journal Article


Abstract


  • Driver cognitive fatigue can significantly affect driving and may lead to fatal accidents. In this regard, automatic detection of underload driver cognitive fatigue based on upper body posture dynamics is studied in this paper, where a semi-supervised approach is developed to identify the cognitive fatigue patterns of driver posture. Initially, an unsupervised Gaussian Mixture Model (GMM) clustering is applied to the acceleration data representing the driver's head, neck, and sternum obtained in a simulated driving through a motion capture suit. This provides the optimum clusters of the most-similar and correlated time-series data of driver upper posture. Then, an automatic labelling algorithm is developed that mines the maximal value and the standard deviation of each GMM cluster and assigns a symbol according to the discrepancy in postural behaviour. Finally, novel machine learning supervised classifiers, including Gaussian Support Vector Machines, and Bootstrap-Aggregating based Ensemble Classifiers, are trained on the GMM-labelled upper body posture dataset, as real-time algorithms, to detect the driver fatigue. The proposed method was validated against cognitive fatigue measured through a neurophysiological method based on an electroencephalogram. The results show that the proposed semi-supervised approach outperforms the existing state-of-art systems in accurately detecting the cognitive fatigue patterns. It successfully recognizes different driving postures with accuracies of 93% and 90% for two test subjects. The shortcomings of the proposed work and directions for potential expansion of current work are discussed.

Publication Date


  • 2022

Citation


  • Ansari, S., Du, H., Naghdy, F., & Stirling, D. (2022). Automatic driver cognitive fatigue detection based on upper body posture variations. Expert Systems with Applications, 203. doi:10.1016/j.eswa.2022.117568

Scopus Eid


  • 2-s2.0-85130155818

Web Of Science Accession Number


Volume


  • 203

Abstract


  • Driver cognitive fatigue can significantly affect driving and may lead to fatal accidents. In this regard, automatic detection of underload driver cognitive fatigue based on upper body posture dynamics is studied in this paper, where a semi-supervised approach is developed to identify the cognitive fatigue patterns of driver posture. Initially, an unsupervised Gaussian Mixture Model (GMM) clustering is applied to the acceleration data representing the driver's head, neck, and sternum obtained in a simulated driving through a motion capture suit. This provides the optimum clusters of the most-similar and correlated time-series data of driver upper posture. Then, an automatic labelling algorithm is developed that mines the maximal value and the standard deviation of each GMM cluster and assigns a symbol according to the discrepancy in postural behaviour. Finally, novel machine learning supervised classifiers, including Gaussian Support Vector Machines, and Bootstrap-Aggregating based Ensemble Classifiers, are trained on the GMM-labelled upper body posture dataset, as real-time algorithms, to detect the driver fatigue. The proposed method was validated against cognitive fatigue measured through a neurophysiological method based on an electroencephalogram. The results show that the proposed semi-supervised approach outperforms the existing state-of-art systems in accurately detecting the cognitive fatigue patterns. It successfully recognizes different driving postures with accuracies of 93% and 90% for two test subjects. The shortcomings of the proposed work and directions for potential expansion of current work are discussed.

Publication Date


  • 2022

Citation


  • Ansari, S., Du, H., Naghdy, F., & Stirling, D. (2022). Automatic driver cognitive fatigue detection based on upper body posture variations. Expert Systems with Applications, 203. doi:10.1016/j.eswa.2022.117568

Scopus Eid


  • 2-s2.0-85130155818

Web Of Science Accession Number


Volume


  • 203