Extended Abstract: Motion Planners Learned from Geometric Hallucination
This addresses the challenge of reducing data or demonstration requirements for robot motion planning, though it appears incremental as it builds on existing planning concepts.
The paper tackles the problem of learning motion planners for robots in obstacle-occupied spaces by introducing a 'hallucination' approach that finds obstacle geometries where a given plan is optimal, and trains an end-to-end planner tested on a physical robot in real-world cluttered environments.
Learning motion planners to move robot from one point to another within an obstacle-occupied space in a collision-free manner requires either an extensive amount of data or high-quality demonstrations. This requirement is caused by the fact that among the variety of maneuvers the robot can perform, it is difficult to find the single optimal plan without many trial-and-error or an expert who is already capable of doing so. However, given a plan performed in obstacle-free space, it is relatively easy to find an obstacle geometry, where this plan is optimal. We consider this "dual" problem of classical motion planning and name this process of finding appropriate obstacle geometry as hallucination. In this work, we present two different approaches to hallucinate (1) the most constrained and (2) a minimal obstacle space where a given plan executed during an exploration phase in a completely safe obstacle-free environment remains optimal. We then train an end-to-end motion planner that can produce motions to move through realistic obstacles during deployment. Both methods are tested on a physical mobile robot in real-world cluttered environments.