Balancing a 3D Inverted Pendulum using Remote Magnetic Manipulation

Jasan Zughaibi, Bradley J. Nelson, Michael Muehlebach

Abstract

Magnetic navigation offers wireless control over magnetic objects, which has important medical applications, such as targeted drug delivery and minimally invasive surgery. Electromagnetic Navigation Systems (eMNSs) are believed to have a superior actuation bandwidth, facilitating trajectory tracking and disturbance rejection. To showcase these dynamic capabilities, we successfully stabilize a non-magnetic inverted pendulum on the tip of a magnetically driven arm. Our approach employs a model-based framework leveraging Lagrangian mechanics and system identification to characterize the system's nonlinearity and bandwidth. A state-feedback controller stabilizes the pendulum, while an iterative learning control (ILC) scheme enables accurate tracking of non-equilibrium trajectories. This work represents the first demonstration of stabilizing a 3D inverted pendulum through electromagnetic navigation.

Introduction: High-Bandwidth Magnetic Navigation

Magnetic navigation systems (MNS) are the future of minimally invasive surgery, allowing doctors to steer catheters or micro-robots through the body using external magnetic fields. However, most current systems are limited by low actuation bandwidth. Systems using permanent magnets moved by robotic arms have high inertia, and many existing models rely on "quasi-static" assumptions that ignore momentum. This is a major hurdle for procedures in dynamic environments, like the cardiovascular system, where a device must compensate for a heartbeat or rapid blood flow.

To prove that Electromagnetic Navigation Systems (eMNS) are up to the task, researchers tackled a classic "boss level" problem in control theory: balancing an inverted pendulum. Because an inverted pendulum is inherently unstable, it requires a high actuation bandwidth—the system must sense a tilt and correct it almost instantly.

Experimental Setup and Methodology

The researchers used the OctoMag, a system featuring eight electromagnets, to control a magnetically driven arm. On top of this arm sits a non-magnetic 3D pendulum, free to rotate in any direction.

The OctoMag system setup featuring eight electromagnetic coils and the inverted pendulum assembly.

Key technical components include:

  • Dynamic Modeling: Instead of simple mapping, the team used Lagrangian mechanics to capture the complex interplay between magnetic torques and the physical momentum of the pendulum. They linearized the 3D dynamics into two decoupled 2D planes to simplify control.
  • System Identification: Frequency-domain analysis was performed to measure the "speed" of the system. The OctoMag demonstrated a bandwidth of ~26.4 Hz, significantly higher than the ~3 Hz required to track a human heartbeat.
  • Hybrid Control Architecture: A state-feedback controller (LQR) keeps the pendulum upright at equilibrium, while Iterative Learning Control (ILC) enables precise movement along a path. ILC analyzes errors from previous iterations of a trajectory (like a figure-eight) and pre-emptively corrects the magnetic field to cancel out repetitive disturbances.

Key Insights: Bandwidth and Scalability

The study highlights that while permanent magnet systems are common, eMNS provides an order-of-magnitude advantage in response time, making it uniquely suited for "active" disturbance rejection in surgery. Furthermore, the researchers confirmed the scalability of this approach by testing a clinical-scale system (Navion), which maintained a 17.3 Hz bandwidth—sufficient for most medical applications.

Results: Precision Stabilization and Tracking

The system successfully balanced pendulums ranging from 20cm to 40cm in length. For context, manually balancing a 20cm pendulum is considered beyond the physical capabilities of most humans.

Experimental results showing the improvement in tracking accuracy for circular and figure-eight trajectories using Iterative Learning Control (ILC).

In tracking maneuvers, the ILC proved essential. Without learning, the system had a tracking error of 1.7°. After just four iterations of the ILC, the error dropped to 0.5°, a 70% improvement. This confirms that eMNS can generate reliable torques at frequencies high enough to handle cardiac interventions, where precise movement is required despite constant physiological motion.

Conclusion and Future Work

This demonstration confirms the feasibility of using electromagnetic navigation for high-speed, dynamic control tasks. While the current setup treats magnetic field gradients as disturbances, future work will focus on using these gradients as additional control inputs. This would allow for even more complex maneuvers, such as controlling the 3D position and orientation of a device simultaneously with even higher precision.

Reference: Zughaibi, J., Nelson, B. J., & Muehlebach, M. (2025). Dynamic Electromagnetic Navigation. IEEE Robotics and Automation Letters.

Summary updated

Abstract

Magnetic navigation offers wireless control over magnetic objects, which has important medical applications, such as targeted drug delivery and minimally invasive surgery. Magnetic navigation systems are categorized into systems using permanent magnets and systems based on electromagnets. Electromagnetic Navigation Systems (eMNSs) are believed to have a superior actuation bandwidth, facilitating trajectory tracking and disturbance rejection. This greatly expands the range of potential medical applications and includes even dynamic environments as encountered in cardiovascular interventions. To showcase the dynamic capabilities of eMNSs, we successfully stabilize a (non-magnetic) inverted pendulum on the tip of a magnetically driven arm. Our approach employs a model-based framework that leverages Lagrangian mechanics to capture the interaction between the mechanical dynamics and the magnetic field. Using system identification, we estimate unknown parameters, the actuation bandwidth, and characterize the system's nonlinearity. To explore the limits of electromagnetic navigation and evaluate its scalability, we characterize the electrical system dynamics and perform reference measurements on a clinical-scale eMNS, affirming that the proposed dynamic control methodologies effectively translate to larger coil configurations. A state-feedback controller stabilizes the inherently unstable pendulum, and an iterative learning control scheme enables accurate tracking of non-equilibrium trajectories. Furthermore, to understand structural limitations of our control strategy, we analyze the influence of magnetic field gradients on the motion of the system. To our knowledge, this is the first demonstration to stabilize a 3D inverted pendulum through electromagnetic navigation.
BibTeX
@article{zughaibi2025balancing,
  author = {Jasan Zughaibi and Bradley J. Nelson and Michael Muehlebach},
  title = {{Balancing a 3D Inverted Pendulum using Remote Magnetic Manipulation}},
  journal = {IEEE Robotics Autom. Lett.},
  year = {2025},
  doi = {10.1109/lra.2025.3563130},
}
Figure