Objectively grade a surgeon through motion and video analysis using statistical modeling techniques.
Skill learning is an emerging area of human-machine collaborative control and automation. As personal robots are increasingly integrated into everyday human life, it will be important to quantify how humans skillfully perform various tasks, to develop methods of assisting humans to learn tasks more efficiently, and then generalizing those
skills for autonomous robot movement. Evaluating skill is a time-consuming, subjective, and difficult process. Robotic minimally invasive surgery (RMIS) has the potential to revolutionize our understanding of modeling, teaching, and evaluating human manipulation skills for data recording without extraneous sensors.
In this talk, I will focus on developing and automating assessment methods for measuring skill and technical competence of human motion. My work has addressed two main aspects of skill: motion modeling through statistical methods and providing augmented feedback to a novice. I examine how to evaluate skill between an expert and novice surgeon, then explore augmented feedback methods to help trainees. I then present a method of automating a surgical robot to perform part of a task using models trained on expert data.
These strategies will improve the performance of trainees by improving the rate of learning. Applications of these techniques for minimally invasive surgery and medical simulation is promising, as it will facilitate new procedures, improve patient outcomes, and reduce training costs. Skill learning can meet the needs of robotics outside the operating room and facilitate adoption in newly identified areas.