Concealment and Obstacle Detection for Autonomous Driving
Tommy Chang, Tsai Hong Hong, Steven Legowik, Marilyn N. Abrams
Obstacle detection and mapping are essential for unmanned autonomous driving. This paper describes both the sensors and the supporting software used in our system for driving autonomously on cross-country roads. The Ladar Range Imaging Camera (EBK) is used for monitoring the environment. We describe an algorithm developed at the National Institute of Standards and Technology (NIST) for detecting obstacles and regions of concealment, and evaluate its ability to detect positive (e.g., rocks) and negative (e.g., ditches) obstacles and concealment regions. We discuss the mapping system used for representing general obstacles (positive, negative, and concealment.) In addition to using information provided by the EBK sensor, the mapping algorithm also uses information supplied by a Global Positioning System (GPS) and an Inertial Navigation System (INS). This system has been tested at NIST and has successfully detected obstacles and regions of concealment while driving cross country at speeds of 35 km/h.
Submitted to the International Association of Science & Technology for Development - Robotics & Applications99 Conference
October 28-30, 1999
Santa Barbara, CA
Science & Technology for Development - Robotics & Applications99 Conference