P-RRAX: A reimplementation of pRRTC through the Jax FFI

P-RRAX is a CUDA-accelerated implementation of the Parallel Rapidly-exploring Random Tree (pRRTC) motion planning algorithm, designed for integration with PyRoFFI's robotics kinematics and collision checking framework. This implementation builds on many pRRTC related work for a robust, batched, and accelerated point-to-point motion validator.

Setup

prrax depends on pyroffi v1.0. PyRoFFI vendors a CUDA backend through GRiD and custom motion planning kernels as well as cricket2 (name is wip) which trace compiles CUDA code.

conda activate prrax
pip install -e ../pyroffi-grim
bash ../pyroffi-grim/build_kernels/build_cricket.sh     # installs cricket2 (GRiM's trace compiler)
ln -sfn ../pyroffi-grim pyroffi

Prerequisites: CUDA toolkit (nvcc), Python 3.11+, jax/jaxlib with CUDA.

The following papers were used as reference material in the implementation of this repository:

References

1. pRRTC — Chih H. Huang, Pranav Jadhav, Brian Plancher, and Zachary Kingston.
   pRRTC: GPU-Parallel RRT-Connect for Fast, Consistent, and Low-Cost Motion Planning.
   IEEE International Conference on Robotics and Automation (ICRA), 2026.
   Paper | Code

2. pAORRTC — Chih H. Huang et al.
   GPU-Parallel Asymptotically Optimal RRT-Connect.
   Workshop presentations, ICRA 2026.
   Project

3. MR. POP — Chih H. Huang, Roy Xing, Brian Plancher, and Zachary Kingston.
   MR. POP: Multi-Robot Parallel Optimizing Planner for Almost-Surely Asymptotically Optimal Planning.
   arXiv preprint, 2026.
   Paper

4. cpRRTC — Jiaming Hu, Jiawei Wang, and Henrik I. Christensen.
   cpRRTC: GPU-Parallel RRT-Connect for Constrained Motion Planning.
   RSS Workshop on Motion Planning and Control via Parallelization, 2025.
   Paper

5. McVAMP — Shrutheesh R. Iyer, I-Chia Chang, Andrew Z. Liu, Yan Gu, and Zachary Kingston.
   Vectorizing Projection in Manifold-Constrained Motion Planning for Real-Time Whole-Body Control.
   IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2026.
   Paper | Project

6. Akima Spline — Hiroshi Akima.
   A New Method of Interpolation and Smooth Curve Fitting Based on Local Procedures.
   Journal of the ACM, 17(4), 589–602, 1970.
   Wikipedia | DOI