A Simulation Framework for Magnetic Continuum Robots

Abstract

Remote magnetic navigation is a technology used to robotically steer magnetic medical instruments, such as magnetic catheters and guidewires, for minimally invasive surgery. The ability to model and simulate the behavior of these magnetic instruments in complex anatomies is important for their clinical use in many ways. Simulation frameworks can improve their design, characterization, and automatic control capabilities, as well as provide training simulators for physicians. In this work we introduce a new simulation framework that accounts for both magnetic actuation and interactions forces with meshed collision models. The simulations are validated experimentally in planar rigid models using a pre-clinical electromagnetic navigation system. We also demonstrate the use of our framework to build training simulators for two endovascular navigation tasks including the exploration of the aortic arch and the internal carotid artery.

Magnetically Actuated Continuum Medical Robots: A Review

“…[135] The configuration of equipping multiple permanent magnets distributed along the tip has also been explored. [106,111,141,142] For instance, Armacost et al proposed . Magnetic guidewires and catheters based on permanent magnets and magnetic particles.…”

“…[ 135 ] The configuration of equipping multiple permanent magnets distributed along the tip has also been explored. [ 106,111,141,142 ] For instance, Armacost et al proposed a three‐magnet‐tip catheter, which had the ability to generate a more significant deflection and a more rapid turning compared with the single‐magnetic design (Figure 7b). [ 106 ] They also derived a comprehensive mathematical model to predict the position and deflection of the catheter, taking both the magnetic torques and forces into consideration.…”

LapGym -- An Open Source Framework for Reinforcement Learning in Robot-Assisted Laparoscopic Surgery

“…This characteristic not only enables rapid iteration and optimization of learning algorithms, it also enables realistic human interaction, which is relevant for recording (expert) demonstrations and validating human-robot collaboration. SOFA is an active open source project that is continuously increasing in popularity, attracting new contributors and accruing new features such as dynamic mesh refinement, control of magnetic continuum robots (Dreyfus et al, 2022), and synthetic data generation from numerical simulations for training machine learning models (Mimesis Inria Research Team, 2022;Linkerhägner et al, 2023).…”

“…The aim of these tools is to provide both experienced SOFA users and newcomers with an easy-to-use framework for RL. To demonstrate sofa env's ease of use, we port the existing SOFA simulation from Dreyfus et al (2022) and define a RL environment for its task. The scene contains a magnetic continuum robot that is controlled through an external force field in order to navigate in a model of (a) an aortic arch and (b) a 2D toy scene, as shown in Figure 3.…”

Electromagnetic Manipulation System for Semi-Autonomous Control of Small-Scale Magnetic Objects With Sequential Programming

“…Most systems with paired coils tend to enclose the workspace via circular [17] or square [15] coils with air cores. In contrast, distributed electromagnet-based systems usually have columnar [18] or rectangular [19] coils positioned around and pointing towards the workspace. For general electromagnets, iron cores made of soft magnetic materials with low hysteresis and high permeability are often installed inside the coils to increase the magnetic field intensity [4], [20].…”