HUMAN-ROBOT COLLABORATION IN ROBOTIC-ASSISTED SURGICAL TRAINING

Abstract

An image guided robot assisted surgical training system is proposed in this thesis. It consists of virtual patient, surgical simulation platform, and a robot to learn and transfer the surgical skills. This thesis focuses on the mechanism of learning and transferring surgical skills. A robot was built to resemble the tools and operating scenario of a laparoscopic surgery. A mean shift based method was proposed to identify the motion primitives in the surgical task. Gaussian Mixture Model and Gaussian Mixture Regression were applied to model the surgical skills and reconstruct a generic model of the skills respectively. Hidden Markov Model was applied to recognition of the intention of a user when he/she was operating on the virtual patient. The proposed training method was evaluated by comparing the performance of two groups of subjects. Technical and Clinical evaluation were conducted. Both studies show that the subjects which underwent the proposed training method performed better.