The world’s aging population will face a potential deficit of more than 40K needed surgeons by the end of year 2020. In cardiothoracic surgery, the workload-per-surgeon is expected to more than double by 2035. AI surgical robotics may help address this issue by providing a set of "helping-hands" to assist doctors during surgery. For this, the robot must co-perform task planning – to determine the high-level surgical procedure – and motion planning – to execute the task plan as a series of precise mechanical movements that effect the surgery.
In this project, we will develop a proof-of-concept surgical task planner, which models procedures as a sequence of human-readable and formal actions that transform the robot's initial state to some surgeon-intended goal state; e.g. a robot’s gripper must be co-located with a desired anatomy to 'grasp' the anatomy. Our challenge is to develop a surgical robot whose formal model aligns with the surgeon’s mental model of the procedure. We will work closely with surgeons to refine our models such that surgical robots act in a way that is interpretable to their human partners.
We will: (1) develop new knowledge in surgical robotics and human-robot interaction, (2) disseminate findings at leading conferences and within university outreach events, (3) afford a Computing student to collaboratively work with surgeons to transform lives, (4) ensure Utah's long-term vitality by developing initial results to secure further extramural funding from the NSF and NIH.
Name: Human-Surgical Robotics
Project Stage: Seedling
Categories: Artificial Intelligence,Medicine
Tags: Surgical Robotics,Surgeon-Robot Collaboration,Shared Autonomy,Task Planning,Classical Planning,Human-Robot Interaction