Enhancing AV Traffic Safety through Pedestrian Detection, Classification, and Communication

Enhancing AV Traffic Safety through Pedestrian Detection, Classification, and Communication

This proposal is the next research step in NCDOT’s support of small autonomous vehicles. Through NCDOT support, the Institute for Transportation Research and Education (ITRE) at North Carolina State University (NCSU) is currently developing and piloting small automated vehicle (AV) technology, referred to as EcoPRT. EcoPRT will employ small autonomous vehicles that will enhance user access and mobility options for navigating the NCSU campus in an efficient and relatively inexpensive manner. The concept is proposed as a remedy to first-mile, last-mile and low density transportation issues in North Carolina on university campuses, military bases, business parks, and downtown areas. Presently, it is likely that EcoPRT will operate on existing surface infrastructure, including sidewalks, multi-use paths, and roadways. A critical traffic safety issue has emerged during the first round of EcoPRT development: while EcoPRT can detect objects, how can it identify and interact with pedestrians, and also bicyclists, on a shared-use path? NCSU engineers have developed sophisticated sensors for the EcoPRT vehicles and will collaborate with researchers from NC Agricultural & Technical State University (NC A&T) to refine the system for accurate identification of obstacles and humans utilizing vision-based classification via deep learning neural networks. This project will focus on developing AV communication with pedestrians and explore the considerations and potential for doing so with bicyclists. The findings from this project will help improve traffic safety and AV integration into North Carolina communities. While AV pilot projects are underway across the globe, few are studying the capability of driverless vehicles to communicate with pedestrians and bicyclists on shared use paths. As AVs are able to discern pedestrians and bicyclists from other obstacles, the EcoPRT team will study and develop visual and auditory methods for communicating bi-directional intent between autonomous vehicles and human road users. The EcoPRT research team is uniquely qualified to pursue this research objective. Dr. Seth Hollar is the current P.I. of the NCDOT-funded EcoPRT pilot project and an expert in autonomous vehicle technology. Abdollah Homaifar, the Director of ACIT institute at NC A&T is an expert in machine learning and AI. Dr. Ali Karimoddini is the Deputy Director of the TECHLAV Center of Excellence in Autonomy at NC A&T, where he is an expert in object detection and classification. Dr. Jing Feng is a subject matter expert in advanced human-machine interactions and human factors. Mathew Palmer, the project’s Principal Investigator, is a Research Scholar at ITRE with over seven years’ experience in pedestrian and bicycle safety and the operations lead for the current EcoPRT pilot project.

• Record URL:
  http://connect.ncdot.gov/projects/research/Pages/ProjectSearch.aspx
• Record URL:
  http://connect.ncdot.gov/projects/research/Pages/ProjDetails.aspx?ProjectID=2019-28

Language

• English

Project

• Status: Completed
• Funding: $192897
• Contract Numbers:

  FHWA/NC/2019-28

• Sponsor Organizations:

  North Carolina Department of Transportation

  Research and Development
  1549 Mail Service Center
  Raleigh, NC  United States  27699-1549
• Managing Organizations:

  North Carolina Department of Transportation

  Research and Development
  1549 Mail Service Center
  Raleigh, NC  United States  27699-1549
• Project Managers:

  Penny, Lisa

• Performing Organizations:

  North Carolina State University

  Institute of Transportation Research and Education
  Raleigh, North Carolina  United States  27695-8601
• Principal Investigators:

  Chase, R

• Start Date: 20170801
• Expected Completion Date: 20200731
• Actual Completion Date: 20210731
• Source Data: 2019-28

Subject/Index Terms

• TRT Terms: Access; Autonomous vehicles; Campus transportation; Campuses; Communication; Detection and identification systems; Machine vision; Mobility; Pedestrian detectors; Pedestrian vehicle interface; Traffic safety
• Identifier Terms: North Carolina Department of Transportation
• Subject Areas: Data and Information Technology; Highways; Pedestrians and Bicyclists; Safety and Human Factors; Vehicles and Equipment;

Filing Info

• Accession Number: 01746003
• Record Type: Research project
• Source Agency: North Carolina Department of Transportation
• Contract Numbers: FHWA/NC/2019-28
• Files: RIP, STATEDOT
• Created Date: Jul 23 2020 1:07PM