Vision-based real-time position control of a semi-automated system for robot-assisted joint fracture surgery

Purpose Joint fracture surgery quality can be improved by robotic system with high-accuracy and high-repeatability fracture fragment manipulation. A new real-time visionbased system for fragment manipulation during robotassisted fracture surgery was developed and tested. Methods The control strategy was accomplished by merging fast open-loop control with vision-based control. This two-phase process is designed to eliminate the open-loop positioning errors by closing the control loop using visual feedback provided by an optical tracking system. Evaluation of the control system accuracy was performed using robot positioning trials, and fracture reduction accuracy was tested in trials on ex vivo porcine model. Results The system resulted in high fracture reduction reliability with a reduction accuracy of 0.09mm (translations) and of 0.15 degrees (rotations), maximum observed errors in the order of 0.12mm (translations) and of 0.18 degrees (rotations), and a reduction repeatability of 0.02mm and 0.03 degrees. Conclusions The proposed vision-based system was shown to be effective and suitable for real joint fracture surgical procedures, contributing a potential improvement of their quality.