Evaluating Autonomous Vehicles' Safety Benefits in Mixed Autonomy Scenarios

Details

• Creators:
  Joe-Wong, Carlee ; Yağan, Osman ; Lin, I-Cheng
• Corporate Creators:
  Carnegie Mellon University ; Carnegie Mellon University. Traffic21 Institute. Safety21 University Transportation Center (UTC)
• Corporate Contributors:
  United States. Department of Transportation. University Transportation Centers (UTC) Program ; United States. Department of Transportation. Office of the Assistant Secretary for Research and Technology
• Subject/TRT Terms:
  • Autonomous vehicles
  • Intelligent vehicles
  • Vehicle safety
  • Transportation safety
  • Driving behavior
  • Pedestrians
  • Mobility
  • Topology
  • Driving simulators
  • Driving
  More +
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States
• Edition:
  Final Report (July 1, 2023-June 30, 2024)
• Corporate Publisher:
  Carnegie Mellon University. Traffic21 Institute. Safety21 University Transportation Center (UTC)
• Abstract:
  Connected autonomous vehicles (CAVs) are gradually advancing towards widespread deployments. CAVs promise to improve transportation safety by operating more efficiently and avoiding incidents like crashes due to human driver error. However, they may cause crashes or other safety incidents themselves, especially when interacting with humans. Our work has three parts: (i) estimating the effective incident rates of CAVs and how they are distributed across a city; (ii) incorporating CAVs' and human drivers' ability to react to human pedestrians; and (iii) evaluating our models and analysis in our mixed-autonomy simulator for a variety of road topologies. The main conclusion of our project is that safety dynamics with CAVs are complex and difficult to predict, requiring sophisticated simulators that are flexible enough to model a range of CAV and human driver behavior. Even complex reinforcement learning models, which can theoretically capture different vehicle objectives and complex decision-making, can struggle to accurately capture vehicle behavior and traffic dynamics, due to the complexity of training such models.
• Format:
  PDF
• Funding:
  69A3552344811
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/77493/dot_77493_DS1.pdf
• File Type:
  [PDF - 1023.43 KB]
• Collection(s):
  Artificial Intelligence Deployment Center of Excellence (AI COE) ; Autonomous Vehicles ; University Transportation Centers
• Main Document Checksum:
  urn:sha-512:38a3d9510d54ab72f2b00a4bff335e67e1492b10188678e96f38e4e94e401f59b50149040001a6f4c66c25cec5b59b7c7d924e454ed622aa5a3968c2f6c6b349