Autonomous Vehicle Challenges for the US Rural Midwest

Details

• Creators:
  Rajamani, Rajesh ; Alai, Hamidreza ; Sharma, Gaurav ; Kyong, Hongjoon ; Pushpalayam, Navaneeth
• Corporate Creators:
  Center for Connected and Automated Transportation (CCAT) ; University of Minnesota
• 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:
  Automated Steering Autonomous Vehicles Data Collection Global Navigation Satellite System (GNSS) Global Positioning System Laser Radar Localization Navigation Systems Road Markings Rural Transportation Satellite Navigation Systems Steering

  More +
• Publication/ Report Number:
  CTS 25-14
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States ; West North Central States
• Edition:
  Final Report (December 2023 – August 2025)
• Corporate Publisher:
  University of Michigan. Center for Connected and Automated Transportation
• Abstract:
  Low volume rural roads present special challenges for AVs, since they can be narrow, often will not have right-sidelane markers, may not even have center lines, and may not be plowed for snow removal. Rural traffic intersections can have missing delineation and signage that are normally provided on higher volume roadways. All of these issues pose major challenges to autonomous driving. This project specifically looks into solutions for localization on rural roads without the use of cameras so that they can work in the absence of lane markers and/or in the presence of snow. The use of RTK-corrected GNSS is explored in detail. A full-scale Chrysler Pacifica MnCAV vehicle at the University of Minnesota was utilized for all data collection. On rural roads without obstructions from trees, RTK-corrected GNSS was found to yield better than 10 cm localization accuracy in terms of standard deviation of error. More than 20 satellites are typically available for use in the real-time position calculation on such roads. However, on roads with dense tree cover, satellite visibility may be completely lost on occasions. The use of Lidar for localization by utilizing an open-source LIO-SAM algorithm was also explored in the project and the results obtained suggest that pre-made maps rather than a SLAM algorithm are needed for reliable localization with Lidar. Since a dual-antenna GNSS system equipped with a high performance IMU can cost well over $15,000, the feasibility of reliably estimating yaw angle and slip angle without the need for an expensive dual-antenna sensor was studied. Two novel approaches, one based on the use of a high gain observer that utilizes IMU and single-antenna GNSS measurements, and another based on the use of a low-cost radar on the car, were developed. Both approaches showed very good performance with yaw angle and slip angle estimates being within approximately 1 degree of ground truth readings. The results obtained in this project indicate that a low-cost single-antenna GNSS system can work reliably for localization on rural roads that have at least some open-sky visibility (don’t have dense tree cover).
• Format:
  PDF
• Funding:
  69A3551747105
• Collection(s):
  University Transportation Centers
• Main Document Checksum:
  urn:sha-512:c8745ebb5d4cd027604943afcd432027e1a44789e1149d44c5931636a6c1c7d32bb21b69fe15fe5968f2530687fa0e341078a16b85fac31c3ccf4a0bcd68c927
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/88289/dot_88289_DS1.pdf
• File Type:
  [PDF - 8.09 MB ]