Traversability Analysis for Unmanned Ground Vehicles

Scientists at Defence R&D Canada – Suffield have been investigating autonomous operation of Unmanned Ground Vehicles (UGVs). In order to navigate unknown terrain effectively, UGVs must be able to create an accurate representation of the operational environment. This is typically done by constructing a geometric representation of the environment, called a Terrain Map, from exteroceptive and proprioceptive data streams. This Terrain Map can further be analyzed to provide a measure of the traversability of the terrain. The resulting Traversability Map can be utilized by path planning and obstacle avoidance algorithms to determine the “best” path to follow. This paper discusses DRDC’s Traversability Map as a method of world representation. The Traversability Map interprets geometric data by calculating statistics about the environment to determine whether an area is traversable or not. In doing so, the Traversability Map interprets geometry from a vehicle specific context, allowing for the unique mobility characteristics of platforms to dictate map parameters.