Skip to content

Differentiable GPU-Parallelized Task and Motion Planning

Authors: William Shen, Caelan Garrett, Nishanth Kumar, Ankit Goyal, Tucker Hermans, Leslie Pack Kaelbling, Tomás Lozano-Pérez, Fabio Ramos

Published: 2024 (Conference Paper)

Source: Robotics: Science and Systems (RSS)

Algorithm: cuTAMP

arXiv: 2411.11833

Summary

Exploits GPU parallelism to simultaneously evaluate thousands of candidate continuous parameter seeds for a given plan skeleton, then applies differentiable gradient-based optimization to each seed in parallel to satisfy the induced continuous constraint satisfaction problem. This combines the discrete search of classical TAMP with massively parallel differentiable optimization, significantly reducing solve times for long-horizon manipulation tasks in highly constrained settings.

Abstract

Planning long-horizon robot manipulation requires making discrete decisions about which objects to interact with and continuous decisions about how to interact with them. A robot planner must select grasps, placements, and motions that are feasible and safe. This class of problems falls under Task and Motion Planning (TAMP) and poses significant computational challenges in terms of algorithm runtime and solution quality, particularly when the solution space is highly constrained. To address these challenges, we propose a new bilevel TAMP algorithm that leverages GPU parallelism to efficiently explore thousands of candidate continuous solutions simultaneously. Our approach uses GPU parallelism to sample an initial batch of solution seeds for a plan skeleton and to apply differentiable optimization on this batch to satisfy plan constraints and minimize solution cost with respect to soft objectives. We demonstrate that our algorithm can effectively solve highly constrained problems with non-convex constraints in just seconds, substantially outperforming serial TAMP approaches, and validate our approach on multiple real-world robots.

Primary

Paper arxiv.org

Standard

arXiv Abstract 2411.11833 arXiv PDF 2411.11833 arXiv HTML 2411.11833

Alternate

cutamp.github.io cutamp.github.io GitHub: NVlabs/cuTAMP github.com research.nvidia.com research.nvidia.com

Tags

  • Task and motion planning

  • TAMP

  • CUDA

  • GPU

  • Parallelized

  • Differentiable optimization

  • Robot manipulation

  • Bilevel planning