Date of Original Version
Abstract or Table of Contents
Planning autonomous long range traverses for planetary exploration and similar robotic missions requires several important capabilities. These include the ability to generate smooth, optimal paths, the ability to reason about constrained path-dependent state variables such as energy, and the ability to replan rapidly in response to new information. Existing path planning approaches provide one or two of these capabilities but fall far short of supporting all three. We present a framed cells approach to path planning which enables the computation of smooth paths in state spaces that include constrained path-dependent variables. The effectiveness of this approach is demonstrated in simulation and on two different robots.