REVIEW

Abstract: At high speeds in natural terrain, unmanned ground vehicles can experience unexpected situations that require rapid hazard avoidance maneuvers. In these scenarios there is often limited time to replan a path based on high-order dynamic models. This paper presents a novel method for high speed navigation and hazard avoidance based on the two dimensional “trajectory space, ” a compact and computationally efficient model-based representation of a robot’s dynamic performance limits on natural terrain. This paper also presents a novel method for trajectory replanning, based on a “curvature matching ” technique. Experimental results on a small high-speed UGV demonstrate the method’s effectiveness. Copyright © 2006 IFA