Autonomous Mobility for the Demo III Experimental Unmanned Vehicles

Abstract

This paper will describe the autonomous mobility controller used under the U.S. Army sponsored Demo III experimental Unmanned Vehicle (XUV) project. The controller is designed for avoiding obstacles and traversing highly unstructured outdoor environments. Behavior generation, world modeling and perception capabilities as well as design architecture will be presented. The architecture used is based on 4D-PCS which uses hierarchical levels of knowledge and decision making. At each successive level of the architecture, planning algorithms generate paths with increasing detail. High level commands developed by the soldiers are automatically refined using 30 meter DTED a priori maps and onboard sensors. The XUVs plan and replan their own routes, speed and gaze. Demo III XUVs have been thoroughly tested on scouting scenarios by the U.S. Army soldiers. The results of this testing will be presented.