Campus researchers will create a center devoted to developing medical robots that can complete repetitive surgical tasks, allowing human surgeons to focus on more complex operations.
The Center for Automation and Learning for Medical Robotics was co-founded by campus electrical engineering and computer sciences professors Ken Goldberg and Pieter Abbeel and postdoctoral researcher Sachin Patil. With funding from the National Science Foundation and two private donors, the center will open in November. The center will focus on robot-assisted surgery, although several other projects include steerable surgical needles and cloud-based collaborative robot learning.
Goldberg, who will direct the center, said the research is opening the door to machine-automated surgery.
“For a long time, surgeons were like pedestrians — they did everything manually,” Goldberg said. “A car made it easier to get around, but it still had a human driver. And that’s what we mean when we say ‘robot.’ ”
Goldberg added that robots can handle “tedious” surgical tasks such as cutting and suturing so that surgeons can be more alert for the critical parts of the operation. He said a surgeon may perform suboptimally while trying to complete a repetitive task.
The center trains a surgical robot, called da Vinci, to complete tasks autonomously. The da Vinci robot currently used in operations is controlled by a human surgeon. Goldberg said a future application of this technology could be completing surgeries remotely, for example in a difficult-to-access developing nation.
Stanford University associate professor of mechanical engineering Allison Okamura, who has collaborated with Goldberg and Abbeel, said the challenge with autonomous robots is ensuring that their behaviors are predictable.
While aspects of procedures involving bone, such as hip or knee replacements, can be performed by the robots without supervision, soft-tissue operations have not been possible. The center released a video in October showing a breakthrough in which the da Vinci robot completes automated tasks on soft-tissue materials after observing human-controlled demonstrations.
Patil, who serves as executive director, notes that it is far easier for robots to handle “rigid” materials than substances with human-tissue-like consistency.
Abbeel said the more “variability” and more “uncertainty” concerning soft, “deformable” tissue in a surgical environment, the more difficult it becomes for them to program a robot to respond. He said that programming a robot to complete specific maneuvers becomes difficult and that teaching the robot to observe the world and adapt its motions accordingly is more effective.
Goldberg hopes the center will allow the research to establish a stronger identity and focus.
“(The center) becomes like a magnet for all kinds of top people, and that’s really our primary goal, because it can accelerate the research, and help us grow,” Goldberg said.