Skip to content

Judo: A User-Friendly Open-Source Package for Sampling-Based Model Predictive Control

Authors: Albert H. Li, Brandon Hung, Aaron D. Ames, Jiuguang Wang, Simon Le Cleac'h, Preston Culbertson

Published: 2025 (Workshop Paper)

Source: Robotics: Science and Systems (RSS)

Algorithm: Judo

arXiv: 2506.17184

Summary

Judo is an open-source Python package providing standardized implementations of sampling-based MPC algorithms (MPPI, CEM, Predictive Sampling, etc.), benchmark tasks, and an interactive GUI for controller tuning. It uses MuJoCo as its physics backend for real-time performance, and supports asynchronous execution to ease sim-to-hardware transfer. The focus is on tooling and usability rather than novel algorithmic contributions. One interesting thing is that Judo runs 100% on CPU (by necessity rather than choice, as the authors indicate they are waiting for development of mujoco_warp to stabilize to provide GPU-based sim rollouts). If you get a server grade Threadripper CPU with 64 cores/128 threads you can do amazing things and plan/control in < 2ms.

Abstract

Recent advancements in parallel simulation and successful robotic applications are spurring a resurgence in sampling-based model predictive control. To build on this progress, however, the robotics community needs common tooling for prototyping, evaluating, and deploying sampling-based controllers. We introduce Judo, a software package designed to address this need. To facilitate rapid prototyping and evaluation, Judo provides robust implementations of common sampling-based MPC algorithms and standardized benchmark tasks. It further emphasizes usability with simple but extensible interfaces for controller and task definitions, asynchronous execution for straightforward simulation-to-hardware transfer, and a highly customizable interactive GUI for tuning controllers interactively. While written in Python, the software leverages MuJoCo as its physics backend to achieve real-time performance, which we validate across both consumer and server-grade hardware. Code at https://github.com/rai-opensource/judo.

Primary

Paper arxiv.org

Standard

arXiv Abstract 2506.17184 arXiv PDF 2506.17184 arXiv HTML 2506.17184

Alternate

rai-inst.com rai-inst.com rai-inst.com rai-inst.com GitHub: bdaiinstitute/judo github.com

Tags

  • Model predictive control

  • Sampling-based control

  • MPPI

  • MuJoCo

  • Open source

  • Software

  • Real-time control