2017 Research Project

Using Advanced Analytics to Frame Vulnerable Road User Scenarios with Autonomous Vehicles

Principal Investigator
Noreen McDonald
University of North Carolina, Chapel Hill
View Bio

Asad J. Khattak
University of Tennessee, Knoxville
View Bio


Connected and automated vehicle (CAV) technologies can dramatically improve safety by reducing human errors, which contribute substantially (an estimated 94 percent) to roadway crashes. CAVs can eventually operate effectively on roadways without experiencing decreased performance due to distraction or fatigue.

However, technological advances will not uniformly decrease crash risks. Some environments, crash types, and user groups will continue to experience elevated risks, particularly vulnerable road users such as pedestrians.

This project will address these critical safety issues by: 

  • Assessing the current and future landscape of pedestrian and vehicle conflicts.
  • Identifying how vehicle technology, planning policies, and data analytics can provide systemic solutions to pedestrian-vehicle conflicts.
  • Using big data analytics from vehicle-to-pedestrian, vehicle-to-vehicle, and vehicle-to-infrastructure communications to identify dangerous pre-crash behaviors.

This trans-disciplinary and multimodal approach is critical because solutions require insights from multiple fields. For example, data science provides ideas on how to use data from CAVs to manage the system and identify conflict situations; travel behavior provides insights into how travel patterns will change in the future and how this will affect risk profiles; and planning offers lessons about how to design transportation infrastructure to reduce risks that technology alone cannot ameliorate.

The specific tasks will include literature reviews on current patterns of pedestrian-vehicle conflicts, assessment of how planning and physical design strategies can reduce pedestrian-CAV conflicts, and an analysis of time use (ATUS) and travel survey data (NHTS) to assess mobility trends. Furthermore, risk analysis will be conducted based on analysis of case studies from Ann Arbor, MI (available through Research Data Exchange) and an assessment of how automated vehicle technology will impact crash risk and potential countermeasures. The team will analyze safety data and propose a framework to link automation technology to human error/crash typologies.

Overall, the study will apply innovative statistical, artificial intelligence, and visualization tools to extract valuable information from data, with the purpose of improving safety across modes, especially for vulnerable road users.

Project Details

Project Type: Research
Project Status: Active
Start Date: 3-1-2017
End Date: 9-30-2018
Contract Year: Year 1
Total Funding from CSCRS: $340,000
Co-sponsors: Tennessee Department of Transportation
Collaborating Organizations: University of North Carolina, Chapel Hill; University of Tennessee, Knoxville