2019 Research Project

Driver impairment detection and safety enhancement through comprehensive volatility analysis

Principal Investigator
Asad J Khattak
University of Tennessee, Knoxville
View Bio

Co-Principal Investigator
Subhadeep Chakraborty
University of Tennessee, Knoxville
View Bio

Co-Principal Investigator
Michael Clamann
University of North Carolina, Chapel Hill
View Bio


Summary

Safe driving is highly correlated with driving behavior, as it has been widely reported that over 90% of crashes are a result of human error [NHTSA 2017]. Impaired driving is a key contributing factor leading to 10,497 fatalities (28% of all transportation crash-related deaths) in 2016 [FHWA 2018]. In recent years, with the ubiquity of sensors and increasing computational resources, it has become possible to monitor driver, vehicle and roadway/environment to extract useful information from multi-dimensional data streams coming in from diverse sources. Through a National Science Foundation study, the team has developed the concept of driving volatility, to quantify variations in vehicle kinematics, driver biometrics, and the roadway environment. The study will explore how to measure driver-vehicle-roadway volatilities using a Naturalistic Driving Study data and driving simulator. By integrating and fusing multiple data sources including driver biometrics, vehicle kinematics, and roadway/environment conditions in real-time, this project aims to generate useful feedback to drivers and warnings to surrounding vehicles regarding hazards. The key objectives of this research are to:

  • Develop 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 will harness the data and quantify variations in driver biometrics and behavior, vehicle kinematics, and roadway/environmental conditions utilizing the concept of volatility.
  • Analyze Naturalistic Driving Study data to explore correlations of driver biometrics, driving style, and roadway/environmental characteristics with driver impairment and crash risk.
  • Using multi-modal-multi-user virtual reality simulator, collect and process driver, vehicle and roadway data. Develop algorithms to identify driving impairment by monitoring data streams emanating from driver, vehicle, and roadway in real-time. Based on algorithms, provide feedback and warnings to the driver and possibly to the surrounding vehicles.

Project Details

Project Type: Research
Project Status: Active
Start Date: 8-1-2019
End Date: 8-1-2020
Contract Year: Year 3
Total Funding from CSCRS: $37,930