Driver impairment detection and safety enhancement through comprehensive volatility analysis

This report documents the activities undertaken by the research team during the first year of the project. Combining the team’s earlier work with new efforts, we have developed a framework for obtaining, processing, and analyzing high-frequency multi-dimensional large-scale data using sensors that monitor the driver, vehicle, and roadways. The framework harnesses the data by exploring volatility. Detailed naturalistic driving study data from the NDS SHRP-2 program was analyzed for obtaining insights on impairment and distracted driving. The risks associated with engagement in non-driving tasks in terms of safety critical events are quantified and discussed. A real-time artificial intelligence method is applied to harness the data and quantify instantaneous crash risk by monitoring driver biometrics (in terms of distraction), vehicular movements, and volatility in driving. The analysis presented can detect anomalies in driving, which can lead to crashes and near-crashes. Finally, the use of experimentation in simulated and naturalistic settings is demonstrated. The entails collection and processing of driver biometric, vehicle, and roadway surroundings data. This effort further includes a review the literature on driver monitoring, as well as setting up the experimentation procedures, which will contribute to future research in driver biometric monitoring and impairment detection.

Publications

• Arvin, R., Kamrani, M., & Khattak, A. J. The role of pre-crash driving instability in contributing to crash intensity using naturalistic driving data. Accident Analysis & Prevention, 132, 105226, 2019.
• Arvin, R., & Khattak, A. J. Driving Impairments and Duration of Distractions: Assessing Crash Risk by Harnessing Microscopic Naturalistic Driving Data. Forthcoming in Accident Analysis & Prevention, 2020.
• Arvin, R., Khattak, A. J. & Qi, H. Crash prediction through unified analysis of driver and vehicle volatilities: Application of 1D-Convolutional Neural Network – Long Short-Term Memory. Under review in Transportation Research Board 100th Annual Meeting for presentation, Washington, D.C., 2020.
• Borhani, S., Arvin, R., Khattak, A., Wang, M., Zhao, X. Predicting Drivers’ Reaction Time in Unexpected Lane Departure Situations Using Brainwave Signals: Application of Machine Learning Techniques. Under review in Transportation Research Board 100th Annual Meeting for presentation, Washington, D.C., 2020.
• Jerome, Z. Arvin, R., & Khattak, A. Why are Most Drivers Not Recognizing Impending Single-Vehicle Collisions and Does this Influence Event Outcomes? Under review in Transportation Research Board 100th Annual Meeting for presentation, Washington, D.C., 2020.