Fast Asymptotically Optimal Kinodynamic Planning via Vectorization

2026-07-04

AI Summary

This work addresses the challenge of achieving both computational efficiency and solution optimality in real-time motion planning for high-dimensional, complex dynamical systems. To this end, we propose PAKR, a parallel motion planner that integrates the AO-x meta-algorithmic framework with vectorized sampling, iterative replanning, and dynamic branch factor control. This approach preserves asymptotic optimality and probabilistic completeness while substantially accelerating computation. Notably, PAKR leverages JAX and the XLA compiler to deliver portable GPU acceleration within a standard Python environment and incorporates MuJoCo-XLA for highly efficient simulation. Experimental results demonstrate that PAKR achieves runtime performance comparable to state-of-the-art GPU-based planners in complex dynamical scenarios, while consistently producing higher-quality solutions and exhibiting strong scalability.

Abstract

Sampling-based motion planners have been shown to be effective for systems with complex kinodynamic constraints and high dimensionality. However, these algorithms struggle to achieve real-time performance, leading to recent efforts to parallelize planning. While GPU-accelerated planners have achieved significant speedups, existing approaches require specialized CUDA programming that limits accessibility and portability. We present Parallel Asymptotically Optimal Kinodynamic RRT (PAKR), a massively parallel kinodynamic planner leveraging JAX and the XLA compiler to achieve GPU acceleration through standard Python tooling. By combining our parallel planner with the AO-x meta-algorithm, we achieve asymptotic optimality through fast iterative replanning. We provide a theoretical analysis of probabilistic completeness, analyze the effects of batch size and branching factor on convergence, and demonstrate scalability to complex dynamics using the MuJoCo-XLA simulator. Experiments show competitive runtimes with state-of-the-art GPU planners and superior solution quality.

Problem

• kinodynamic planning
• real-time performance
• GPU acceleration
• sampling-based motion planning
• portability

Innovation

• kinodynamic planning
• GPU acceleration
• JAX
• asymptotic optimality
• sampling-based planning