Real-time simulation enabled navigation control of magnetic soft continuum robots in confined lumens

Dezhong Tong, Zhuonan Hao, Jiyu Li, Boxi Sun, Mingchao Liu, Liu Wang, Weicheng Huang

TL;DR

This paper introduces a novel framework for real-time navigation control of magnetic soft continuum robots (MSCRs) in confined lumens. By combining a discrete elastic rod model with an efficient contact handling strategy and a model-based control approach, they achieve precise, safe, and autonomous steering of MSCRs, overcoming limitations of existing methods.

What Problem Does This Paper Address?

Minimally invasive surgeries, particularly in vascular and gastrointestinal applications, often rely on guidewires. However, traditional guidewires with pre-shaped tips struggle with navigating complex, tortuous lumens, leading to increased procedure times, risk of vessel injury, and prolonged radiation exposure for patients and surgeons. Magnetic soft continuum robots (MSCRs) offer a promising alternative due to their inherent flexibility and ability to bend under external magnetic fields. While significant progress has been made in modeling MSCR behavior, achieving real-time navigation control remains a major challenge. This is primarily due to the computational expense of simulating MSCRs' nonlinear dynamics and complex contact interactions with lumen walls, making existing methods impractical for clinical use.

Methods

The authors propose a framework that integrates several key components:

  • Hard-Magnetic Elastic Rod Theory with Discrete Differential Geometry (DDG): They model the MSCR as a 3D hard-magnetic elastic rod, capturing large deformations and complex interactions using the Discrete Elastic Rod (DER) algorithm. This approach discretizes the robot into nodes and edges, allowing for efficient computation of its mechanical response.
  • Order-Reduced Contact Handling: To address the computational burden of contact, they employ a penalty energy-based approach. This method penalizes deviations from a "virtual tube" around the lumen centerline, effectively guiding the MSCR away from contact without the need for explicit, computationally intensive contact detection for every point.
  • Model-Based Control Framework: The core of their control strategy is formulated as an inverse design problem. They aim to compute the optimal actuation magnetic field that minimizes the constraint force (i.e., contact force) while ensuring the MSCR tip follows the lumen centerline. This is achieved by embedding the optimization objective within the robot's equilibrium equations and using a Newton-based method to iteratively update the magnetic field.

Datasets

Not discussed in the paper. The authors rely on numerical simulations and experimental validation using custom-built lumen phantoms.

Key Insights and Ideas

  • Bridging Simulation and Control: The paper successfully demonstrates that a computationally efficient, reduced-order simulation model can be directly integrated into a real-time control loop for MSCRs.
  • Contact-Aware Control is Crucial: The results highlight that passive pushing of MSCRs leads to significant contact forces and trajectory deviations, underscoring the necessity of active magnetic control for safe navigation.
  • Nonlinearity is Key: The optimal magnetic field required for navigation exhibits extreme nonlinearity, even in simple geometries. This is attributed to the complex interplay of the MSCR's elasticity, magnetic torques, and contact interactions.
  • Real-time Performance Achieved: The proposed framework achieves real-time performance, making it suitable for dynamic clinical applications.

Key Results

  • Accurate Simulation: The numerical simulations show excellent agreement with experimental results for MSCR navigation in various lumen geometries (half-circle, sinusoidal, helical, and multi-bifurcated).
  • Contact Force Elimination: The model-based control method effectively eliminates distal contact forces, reducing the risk of lumen injury. For instance, in a sinusoidal lumen, passive pushing resulted in contact forces up to 40, while active control reduced them to near zero.
  • Precise Navigation: The MSCR tip accurately follows the lumen centerline, even in complex 3D helical and bifurcated phantoms, demonstrating the versatility of the control strategy.
  • Computational Efficiency: The framework achieves real-time performance with large time steps (dt > 0.003 s) and small optimization steps (δX ≤ 0.1), making it practical for clinical use.

Limitations and Future Work

  • Limitations:
    • The current model assumes a fixed lumen centerline.
  • The control strategy relies on a simplified penalty-based energy formulation for contact.
  • While validated experimentally, the current experimental setup is a simplified representation of real-world clinical scenarios.
  • Future Work:
    • Integrate the methodology into robotic systems for full automation of MSCR applications.
  • Explore more advanced penalty-based energy formulations for optimal control problems.
  • Investigate the predictive power of classical mechanics for controller design in soft matter manipulation.

Summary updated

BibTeX
@article{tong2025real,
  author = {Dezhong Tong and Zhuonan Hao and Jiyu Li and Boxi Sun and Mingchao Liu and Liu Wang and Weicheng Huang},
  title = {{Real-time simulation enabled navigation control of magnetic soft continuum robots in confined lumens}},
  journal = {Journal of the Mechanics and Physics of Solids},
  year = {2025},
  url = {https://openreview.net/forum?id=qP5nZ3F2jJ},
}
Figure