Fast Feature Detection and Stochastic Parameter Estimation of Road Shape using Multiple LIDAR

This paper describes an algorithm for an autonomous car to identify the shape of a roadway by detecting geometric features via LIDAR. The data from multiple LIDAR are fused together to detect both obstacles as well as geometric features such as curbs, berms, and shoulders. These features identify the boundaries of the roadway and are used by a stochastic state estimator to identify the most likely road shape. This algorithm has been used successfully to allow an autonomous car to drive on paved roadways as well as on offroad trails without requiring different sets of parameters for the different domains.

• Record URL:
  http://www.ri.cmu.edu/pub_files/2008/9/peterson_kevin_2008_1.pdf
• 
• Corporate Authors:

  Carnegie Mellon University

  Robotics Institute, 5000 Forbes Avenue
  Pittsburgh, PA  United States  15213-3890

  Defense Advanced Research Projects Agency

  3701 North Fairfax Drive
  Arlington, VA  United States  22203-1714
• Authors:
  • Peterson, Kevin
  • Ziglar, Jason
  • Rybski, Paul E
• Publication Date: 2012

Language

• English

Media Info

• Media Type: Web
• Features: Figures; References;
• Pagination: 8p

Subject/Index Terms

• TRT Terms: Automobile navigation systems; Autonomous vehicle guidance; Geometric configurations and shapes; Laser radar; Stochastic processes
• Subject Areas: Highways; Vehicles and Equipment; I91: Vehicle Design and Safety;

Filing Info

• Accession Number: 01444365
• Record Type: Publication
• Files: TRIS
• Created Date: Aug 28 2012 8:58AM