Vision-Based 3-D Control of Magnetically Actuated Catheter Using BigMag - An Array of Mobile Electromagnetic Coils

Automated steering of endovascular catheters has a potential of improving the outcome of minimally invasive surgical procedures. Nevertheless, actuation, tracking, and closed-loop position control of catheters remain a challenge. In this study, we present a modular framework for a three-dimensional (3-D) position control of magnetically actuated endovascular catheter. The catheter is fitted with a permanent magnet and deflected using externally generated magnetic field provided by BigMag - An array of mobile electromagnets. Pseudorigid-body modeling is used to formulate an inverse-model closed-loop position controller of the catheter. The shape feedback is reconstructed from a 3-D point cloud of catheter silhouette, obtained using stereo vision. Magnetic actuation is enabled using an inverse field map technique, mapping the reference magnetic field to BigMag configuration variables. The framework is tested in a series of experiments. The inverse map is validated, showing a mean magnetic field error of 2.20%. The accuracy of the shape reconstruction algorithm is 0.59 mm. Finally, the magnetically actuated catheter is steered across a series of trajectories with maximum reported catheter deflection of 68.43-\circ and maximum tip speed of 5 mm/s. Across all trajectories, the best control performance metrics are the mean error of 0.57 mm and the RMS error of 0.77 mm