Model-free and Vision-based Contact Force Control of Interventional Catheters

Catheter-tissue contact force(CF) has been reported as one of the success factors of heart ablation surgery. Monitoring and controlling the CF are challenging processes due to the limitation of sensing technologies and robotic surgery systems. In the present study, an image-based and sensor-less CF controller system is proposed that benefits from a neural force feedback estimator. Firstly, a real-time and accurate force feedback element was designed and evaluated to estimate the CF directly from the image of the ablation catheter’s deflections during the surgery. This learning-based force estimator successfully could provide CF with a mean absolute error of 0.012 ± 0.01 N and it was implemented in a closed-loop controller as a feedback element. Then, an experimental setup including the ablation catheter manipulator mechanism and a camera was designed to evaluate the performance of the CF controller for static and dynamic input CFs. The evaluation tests showed that the system can maintain the CF with root-mean-squared(RMS) error of 0.01 ± 0.01 N for static test and RMS error of 0.04 ± 0.03 N for a dynamic input force.