A novel vision guided knot-tying method for autonomous robotic surgery

Abstract

This paper presents a vision guided automatic knot-tying method for Robotic Assisted Minimally Invasive Surgery. By utilizing 3D position reconstruction of selection points on images from calibrated stereo cameras and coordinating two robot motions for proper suture handling, we successfully realize an automatic knot-tying procedure. Our experimental results show that autonomous knot-tying has the potential to be faster than human performance, thus supporting supervisory control.