A New Method for Classification of Hazardous Driver States Based on Vehicle Kinematics and Physiological Signals
Hazardous driver states are the cause of many traffic accidents, and there is therefore a great need for accurate detection of such states. This study proposes a new classification method that is evaluated on a previously collected driving dataset that includes combinations of four causes of hazardous driver states: drowsiness, high traffic density, adverse weather, and cell phone usage. The previous study consisted of four sessions and eight scenarios within each session. Four physiological signals (e.g. electrocardiogram) and eight vehicle kinematics signals (e.g. throttle, road offset) were recorded during each scenario. In both previous and present studies, the presence or absence of the different causes of hazardous driver states was classified. In this study, a new classifier based on principal component analysis and artificial neural networks is proposed. The obtained results show improvement across all classification accuracies, especially when only vehicle kinematics data are used (mean of 12.7%).
KeywordsHazardous driving state Artificial intelligence Driving performance Physiological measurements Affective computing
Research supported by the National Science Foundation under grant no. 1717705 as well as by the National Institute of General Medical Sciences of the National Institutes of Health under grant no. P20GM103432.
- 1.National Highway Traffic Safety Administration (NHTSA): 2015 Motor Vehicle Crashes: Overview (2016)Google Scholar
- 7.Chuang, C.-H., Cao, Z., King, J.-T., Wu, B.-S., Wang, Y.-K., Lin, C.-T.: Brain electrodynamic and hemodynamic signatures against fatigue during driving. Front. Neurosci. 12 (2018)Google Scholar
- 11.Darzi, A., Gaweesh, S.M., Ahmed, M.M., Novak, D.: Identifying the causes of drivers’ hazardous states using driver characteristics, vehicle kinematics, and physiological measurements. Front. Neurosci. 12 (2018)Google Scholar
- 12.Darzi, A., Gorsic, M., Novak, D.: Difficulty adaptation in a competitive arm rehabilitation game using real-time control of arm electromyogram and respiration. In: 2017 International Conference on Rehabilitation Robotics (ICORR), pp. 857–862. IEEE (2017)Google Scholar
- 15.Zheng, Y., Hansen, J.H.L.: Lane-change detection from steering signal using spectral segmentation and learning-based classification. IEEE Trans. Intell. Veh. 8858, 1 (2017)Google Scholar
- 17.Kordani, A.A., Molan, A.M., Monajjem, S.: New formulas of side friction factor based on three-dimensional model in horizontal curves for various vehicles. In: T&DI Congress 2014, pp. 592–601. American Society of Civil Engineers, Reston (2014)Google Scholar