Sensor-Free Strategy for Estimating Guidewire/Catheter Shape and Contact Force in Endovascular Interventions

Naner Li,Yiwei Wang,Huan Zhao,Han Ding

DOI: https://doi.org/10.1109/lra.2024.3504236

IF: 5.2

2024-01-01

IEEE Robotics and Automation Letters

Abstract:Accurate assessment of guidewire shape and contact forces is critical for autonomous robotic endovascular procedures. However, existing sensor-based approaches often require modifications to standard guidewires or the use of custom-made alternatives, which can hinder integration into conventional surgical workflows and increase costs. Moreover, the sensor-based method can only obtain partial force information. This letter aimed to develop a novel sensor-free two-step computational method for estimating overall guidewire shape and forces using only routinely obtainable information. The vascular space is discretized into multiple mesh points along the centerline to form a graph. Based on an energy equation, the lowest energy path between the start point (the insertion position) and the endpoint (the guidewire tip position) is searched as the initial shape of the guidewire. This initial shape, along with the known insertion length, is then input into a finite element model to compute the final guidewire configuration and contact forces. The method was validated through 3 rounds of testing in 3 phantom models at 4 different insertion lengths. In 11 successful experimental scenarios, the estimated guidewire shapes closely matched the actual shapes, with an average root mean square error of 0.50 $\pm$ 0.12 mm. The contact force estimation achieved an average accuracy of 91.9 $\pm$ 2.9%, with an average angular deviation of 1.94 $\pm$ 1.03 $^\circ$ from measured values.