3-D perception and modeling

Abstract

Autonomous robots are increasingly suited for operations in unstructured application domains. One impressive example is the result of the Defense Advanced Research Projects Agency (DARPA) Grand Challenge 2005, where five teams managed to get autonomous vehicles to complete a 213-km track in the Mojave desert. Some conditions of the Grand Challenge do not hold for space robotics, especially the availability of a drivable track defined by global positioning system (GPS). In general, it can be argued that space robotics and especially planetary exploration have to deal with one of the most challenging unstructured environments. There is absolutely no infrastructure, the cost of mission failure is very high, and many core sensors that form the foundations of intelligent autonomous behaviors are not space proof, yet. However, autonomy is not a binary all or nothing property. Intelligent autonomous functions can supplement different levels of teleoperation. In the context of the 2008 Lunar Robotics Challenge (LRC) of the European Space Agency (ESA), the Jacobs Robotics team investigated three-dimensional (3-D) perception and modeling as an important basis of autonomy in unstructured domains. Concretely, the efficient modeling of the terrain via a 3-D laser range finder (LRF) is addressed. The underlying fast extraction of planar surface patches can be used to improve situational awareness of an operator or for path planning.