MagBotSim: Physics-Based Simulation and Reinforcement Learning Environments for Magnetic Robotics

TL;DR

MagBotSim introduces a physics-based, MuJoCo-backed simulator and RL environments for Magnetic Levitation robotics, enabling high-level motion planning and object manipulation within a scalable, swarming mover-based paradigm. It provides XML-driven custom environment generation, compatibility with Gymnasium and PettingZoo, and benchmark suites for object pushing and trajectory planning to enable fair comparisons across methods. Experimental results demonstrate scalability to large mover swarms on consumer hardware, effective object manipulation tasks, and successful sim2real transfer to a Beckhoff XPlanar system, underscoring practical applicability. The work highlights future enhancements, including magnetic-field modeling, hardware acceleration, and collaborative multi-mover manipulation, to further bridge simulation and industrial deployment in Magnetic Robotics.

Abstract

Magnetic levitation is about to revolutionize in-machine material flow in industrial automation. Such systems are flexibly configurable and can include a large number of independently actuated shuttles (movers) that dynamically rebalance production capacity. Beyond their capabilities for dynamic transportation, these systems possess the inherent yet unexploited potential to perform manipulation. By merging the fields of transportation and manipulation into a coordinated swarm of magnetic robots (MagBots), we enable manufacturing systems to achieve significantly higher efficiency, adaptability, and compactness. To support the development of intelligent algorithms for magnetic levitation systems, we introduce MagBotSim (Magnetic Robotics Simulation): a physics-based simulation for magnetic levitation systems. By framing magnetic levitation systems as robot swarms and providing a dedicated simulation, this work lays the foundation for next generation manufacturing systems powered by Magnetic Robotics. MagBotSim's documentation, videos, experiments, and code are available at: https://ubi-coro.github.io/MagBotSim/

Figures (7)

• Figure 1: Mover on a magnetic levitation system
• Figure 2: Schematic overview of the proposed MagBotSim library, a physics-based simulation for Magnetic Robotics. MagBotSim includes reinforcement learning environments for trajectory planning and object manipulation.
• Figure 3: MagBotSim is designed to create customized environments that match real-world applications.
• Figure 4: Benchmark environments included in MagBotSim.
• Figure 5: On a laptop CPU, it takes about $25\,$ms to execute one simulation step (apply controls, one MuJoCo integrator step, and collision checking) with 1k movers.