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 samplingbased 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/bdaiinstitute/judo.
The paper introduces 'judo', an open-source software package for sampling-based Model Predictive Control (MPC), aimed at the robotics community. The tool provides robust implementations of various MPC algorithms and standard benchmark tasks, facilitating rapid prototyping and evaluation. Key features include a user-friendly interface, asynchronous execution for easy simulation-to-hardware transfer, and a customizable GUI for interactive controller tuning. It is built in Python and utilizes MuJoCo for real-time performance, validated across consumer and server hardware. The software emphasizes extensibility and ease of use compared to existing tools. The paper discusses design principles, core components of the judo package, including the controller interface, GUI functionalities, and asynchronous operation, with evaluations showcasing its performance and capabilities in real-time control scenarios.
This paper employs the following methods:
- Sampling-based MPC
- Predictive Sampling
- Cross-Entropy Method (CEM)
- Model Predictive Path Integral Control (MPPI)
The following datasets were used in this research:
- judo facilitates rapid prototyping and evaluation of sampling-based MPC algorithms.
- The performance of judo is validated across both consumer-grade and server-grade hardware.
The authors identified the following limitations:
- Disparity in performance between consumer-grade and server-grade hardware.
- Number of GPUs: None specified
- GPU Type: None specified
- Compute Requirements: None specified