Two Approaches to Dynamic Refinement in Hierarchical Motion Planning
Tichakorn Wongpiromsarn, Venkatesh G. Rao, Raffaello D’Andrea · AIAA Guidance, Navigation, and Control Conference and Exhibit · 2005
A common technique in robotic motion planning is to progressively refine trajectories using a hierarchical planner comprising several iteratively coupled layers. This paper presents two dynamic refinement techniques that transform the polygonal path computed by a geometric planning layer into a feasible trajectory for a specific vehicle by fitting suitable geometric primitives. This work advances the state of the art in dynamic refinement in two ways. First, we generalize the popular circular-arc-primitive methods for car-like vehicles to achieve aggressive maneuvering and higher speeds. Second, we present a preliminary analysis of minimum-time primitives for omni-directional vehicles. Results of extensive simulation experiments are presented, to demonstrate the performance characteristics of the two methods.