5G-enabled Safe and Robust Deep Multi-agent Reinforcement Learning Framework for CAV Coordination

The project addresses the critical need for reliable and safe deployment of connected and autonomous vehicles (CAVs) using advanced V2X (vehicle-to-everything) communication technologies. The focus is on exploring the integration of 5G-enabled V2X infrastructure with AI-based decision-making and control in CAVs. The primary objective is to enhance the understanding of the relationship between communication quality and CAV safety and efficiency, particularly under complex driving scenarios. The project proposes a novel multi-agent reinforcement learning (MARL) framework that utilizes enhanced observation capabilities from 5G-based V2X communication. This framework aims to ensure rigorous safety for CAVs and is robust across various driving scenarios. It includes the development of real-time traffic scheduling for 5G V2X communication and an evaluation of the effectiveness of these technologies in improving system safety and efficiency. The project involves collaborative research and testbed development, including simulations and small-scale CAV experiments, to validate the safety and efficiency of the proposed methodology. This interdisciplinary research promises significant advancements in CAV communication, machine learning, and control, paving the way for safer and more efficient autonomous driving experiences.

Language

• English

Project

• Status: Active
• Funding: $230000
• Contract Numbers:

  69A3552348301

• Sponsor Organizations:

  Office of the Assistant Secretary for Research and Technology

  University Transportation Centers Program
  Department of Transportation
  Washington, DC  United States  20590
• Managing Organizations:

  University of Massachusetts, Amherst

  Department of Civil and Environmental Engineering
  130 Natural Resources Road
  Amherst, MA  United States  01003
• Performing Organizations:

  University of Connecticut, Storrs

  Connecticut Transportation Institute
  270 Middle Turnpike, Unit 5202
  Storrs, CT  United States  06269-5202
• Principal Investigators:

  Miao, Fei

  Han, Song

• Start Date: 20240101
• Expected Completion Date: 20241231
• Actual Completion Date: 0
• USDOT Program: University Transportation Centers Program
• Subprogram: University Transportation Centers

Subject/Index Terms

• TRT Terms: Autonomous vehicles; Connected vehicles; Machine learning; Multi-agent systems; Vehicle to infrastructure communications; Vehicle to vehicle communications
• Subject Areas: Data and Information Technology; Operations and Traffic Management; Safety and Human Factors; Vehicles and Equipment;

Filing Info

• Accession Number: 01904668
• Record Type: Research project
• Source Agency: New England University Transportation Center
• Contract Numbers: 69A3552348301
• Files: UTC, RIP
• Created Date: Jan 16 2024 12:08PM