Generalized, On-Line, Self-Supervised Learning for Autonomous Vehicles – Looking Farther Ahead using Inferred Geometry

PDF

Authors
  1. Broten, G.S.
  2. Mackay, D.J.
Corporate Authors
Defence R&D Canada - Suffield, Ralston ALTA (CAN)
Abstract
The Autonomous Intelligent Systems Section at Defence R&D Canada – Suffield develops autonomous capabilities for land and air vehicles. Unmanned ground vehicle (UGV) autonomy demands numerous capabilities, including traversing outdoor, unstructured environments. As a rule outdoor terrain is not static as it varies on a seasonal basis due to the life cycle associated with natural flora. Additionally, outdoor terrain may change appearance due to variations in lighting conditions that result from the Sun’s relative position and from weather conditions such as clouds, fog or rain. Finally, unstructured terrain contains many natural obstacles, in addition to rocks and berms, that are impediments to traversal, such as brush, low lying wet areas, and other, seasonally varying, soft soil conditions. The tremendous diversity associated with outdoor terrain has long caused researchers considerable grief. It has made the development of terrain classification algorithms a very difficult if not impossible task. Traditionally, researchers have avoided this problem by relying upon ranging sensors, which provide the data required to construct 2½ D or 3D world representations. Although geometrical representations have been used extensively, the low data rates associated with laser rangefinders, their limited ranges and the shallow angles at which they operate all combine to limit their effectiveness. The principal alternative to active ranging sensors in the visual part of the spectrum i

Il y a un résumé en français ici.

Report Number
DRDC-SUFFIELD-TR-2009-220 — Technical Report
Date of publication
01 Dec 2009
Number of Pages
84
DSTKIM No
CA034489
CANDIS No
533944
Format(s):
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: