Learning Sampling Distributions for Robot Motion Planning

Abstract

A defining feature of sampling-based motion planning is the reliance on an implicit representation of the state space, which is enabled by a set of probing samples. Traditionally, these samples are drawn either probabilistically or deterministically to uniformly cover the state space. Yet, the motion of many robotic systems is often restricted to "small" regions of the state space, due to, for example, differential constraints or collision-avoidance constraints. To accelerate the planning process, it is thus desirable to devise non-uniform sampling strategies that favor sampling in those regions where an optimal solution might lie. This paper proposes a methodology for non-uniform sampling, whereby a sampling distribution is learned from demonstrations, and then used to bias sampling. The sampling distribution is computed through a conditional variational autoencoder, allowing sample generation from the latent space conditioned on the specific planning problem. This methodology is general, can be used in combination with any samplingbased planner, and can effectively exploit the underlying structure of a planning problem while maintaining the theoretical guarantees of sampling-based approaches. Specifically, on several planning problems, the proposed methodology is shown to effectively learn representations for the relevant regions of the state space, resulting in an order of magnitude improvement in terms of success rate and convergence to the optimal cost. * These authors contributed equally to this work.This work was originally presented at the 2018 IEEE International Conference on Robotics and Automation (ICRA). This extended version includes new numerical experiments demonstrating generalization, iterative retraining of the generative model, and multirobot planning experiments.

Tethered Tool Manipulation Planning With Cable Maneuvering

“…The algorithm is used to plan for actions that move user-specified objects in the environment to user-defined locations. [10] presents a methodology for nonuniform sampling to accelerate samplingbased motion planning. In [11], a discrete RRT algorithm for path planning is shown.…”

Learning a Lattice Planner Control Set for Autonomous Vehicles

“…In previous work, data-driven motion planning has often focused on learning search heuristics or policies for the motion planner rather than learning the underlying structure of the planner itself. Ichter et al developed a method for learning a sampling distribution for RRT* motion planning [8]. Imitation learning can also be used to learn a search heuristic based on previously planned optimal paths [9], [10].…”

Decentralized Cooperative Planning for Automated Vehicles with Continuous Monte Carlo Tree Search

“…Related work by Paxton et al integrate a learned highlevel options policy in combination with a low-level control policy into the MCTS to improve the overall quality for a single agent planning problem. [16] Other approaches to accelerate sampling based motion planning task have been proposed, using conditional variational autoencoder, generating subspaces for sampling distributions over desired states [18]. Similarly Banzhaf et al learn a sampling distribution for poses of an RRT* path planning algorithm in semi-structured environments.…”