cuRobo: Parallelized Collision-Free Minimum-Jerk Robot Motion Generation
Abstract¶
This paper explores the problem of collision-free motion generation for manipulators by formulating it as a global motion optimization problem. We develop a parallel optimization technique to solve this problem and demonstrate its effectiveness on massively parallel GPUs. We show that combining simple optimization techniques with many parallel seeds leads to solving difficult motion generation problems within 53ms on average, 62x faster than SOTA trajectory optimization methods. We achieve SOTA performance by combining L-BFGS step direction estimation with a novel parallel noisy line search scheme and a particle-based optimization solver. To further aid trajectory optimization, we develop a parallel geometric planner that is atleast 28x faster than SOTA RRTConnect implementations. We also introduce a collision-free IK solver that can solve over 9000 queries/s. We are releasing our GPU accelerated library CuRobo that contains core components for robot motion generation. Additional details are available at sites.google.com/nvidia.com/curobo.
Summary¶
CUDA-accelerated library for collision-free robot motion generation. Formulates trajectory generation as a global optimization problem solved across thousands of parallel seeds on GPU. Combines L-BFGS with a novel parallel noisy line search and particle-based optimization to produce minimum-jerk, collision-free trajectories within ~50ms. Also includes a parallel geometric planner (~20ms) and a batched IK solver (>7000 queries/s). An earlier version without minimum-jerk optimization was published at ICRA 2023.