Remote Magnetic Levitation Using Reduced Attitude Control and Parametric Field Models

Neelaksh Singh, Jasan Zughaibi, Denis von Arx, Bradley J. Nelson, Michael Muehlebach

Introduction

Electromagnetic Navigation Systems (eMNS) represent a frontier in minimally invasive medicine, offering the potential to guide micro-robots or catheters through the human body with high precision. However, most current applications are "quasi-static," relying on fluid friction (e.g., in blood vessels) for stability. Achieving stable, remote magnetic levitation in air is significantly more challenging due to Earnshaw’s Theorem, which states that static magnetic fields cannot stably trap a magnetic object. This paper presents a framework for achieving stable 5-degree-of-freedom (DOF) levitation of a centimeter-scale object using a clinical-grade eMNS.

The experimental setup showing the 3D-printed levitator suspended within the OctoMag electromagnetic navigation system.

System Architecture and Modeling

The researchers utilized the OctoMag, an eMNS featuring eight electromagnets, to control a 32.4g 3D-printed levitator containing two permanent magnets. To ensure the system is "levitator-agnostic," the team implemented two core innovations:

  1. Parametric Field Modeling: Instead of relying on computationally expensive simulations or large lookup tables, the system uses a Multipole Expansion Model (MPEM). This analytical model maps coil currents to magnetic fields and gradients, allowing the system to be easily adapted to different levitated objects by simply updating dipole strength and mass parameters.
  2. Reduced Attitude Control: A single magnetic dipole cannot be torqued around its own axis, making "yaw" inherently uncontrollable. The researchers developed a nonlinear reduced attitude controller that stabilizes the two controllable rotational degrees of freedom (roll and pitch) while allowing the uncontrollable yaw to drift, preventing the controller from attempting physically impossible corrections.

Experimental Results

The system was tested at a high-speed control loop of 1 kHz, achieving a total system latency of approximately 4 ms. This high bandwidth is critical for counteracting the inherent instability of the magnetic field.

Experimental data showing the levitator tracking sinusoidal trajectories in X and Y while maintaining a constant Z height. The plots demonstrate precise tracking of roll (ϕ) and pitch (θ) setpoints, while the uncontrollable yaw (ψ) is allowed to drift.

Key performance metrics included:

  • Precision Tracking: Successful execution of complex 3D trajectories, such as "figure-eight" maneuvers, with minimal error.
  • High-Angle Stability: The levitator maintained stability during roll and pitch maneuvers of up to ±45°.
  • Disturbance Rejection: Integral control effectively eliminated steady-state errors arising from gravity and minor model inaccuracies.

Discussion and Future Work

While the current system demonstrates robust remote levitation, it currently depends on an external infrared motion capture system for feedback. Transitioning to clinical environments will require integrating medical imaging modalities like X-ray or ultrasound for localization. Additionally, future designs may incorporate multiple dipoles to achieve full 6-DOF control, addressing the current limitation of uncontrollable yaw drift. The ultimate goal is to scale these models for human-scale systems, bringing remote magnetic navigation closer to real-world surgical applications.

Summary updated

BibTeX
@inproceedings{singh2025remote,
  author = {Neelaksh Singh and Jasan Zughaibi and Denis von Arx and Bradley J. Nelson and Michael Muehlebach},
  title = {{Remote Magnetic Levitation Using Reduced Attitude Control and Parametric Field Models}},
  booktitle = {Repository},
  year = {2025},
  doi = {10.3929/ethz-c-000791286},
}
Figure