by A robot carried out realistic surgery for the first time without human help.
Researchers at Johns Hopkins University said on Wednesday that the robot “unshakable” carried out a long phase of a gallbladder in a lifelike patient and reacted to the team’s voice commands. It showed the know-how of a qualified human surgeon, even in unexpected scenarios that are typical of medical emergencies in real life.
“This progress moves us from robots who can carry out certain surgical tasks for robots who really understand the surgical interventions,” said medical robotic Axel Krieger in a statement. “This is a critical distinction that brings us much closer to the clinically sustainable autonomous surgical systems that can work in chaotic, unpredictable reality of actual patient care.”
The robot, which is known as the “SRT-H” or “Surgical Robot Transformer Hierarchy”, had trained on videos of operations and observed John’s Hopkins surgeon, which carried out the process on pork gaps. The videos had captivates that described the tasks.
After the robot had seen the videos, he carried out the operation with an accuracy of 100 percent, said John’s Hopkins. Although the robot took longer to carry out the operation, the results were comparable to that of a surgeon.
The robot, which was built with the same mechanical learning architecture, that provides the popular chatbot chat-chat chats in artificial intelligence, had to do 17 minutes long tasks, including the identification of channels and arteries, to take them precisely, strategically climb and drain parts with scissors.
This is more than as a warrior “Smart Tissue Autonomous Robot” three years ago, the first autonomous robot surgery over an animal – a laparoscopic operation led over a living pig. This robot also needed specially marked tissue, operated in a heavily controlled environment and followed a rigid surgical plan.
Conversely, SRT-H made perfectly over anatomical conditions that were not uniform. This was even the case when the researchers changed the output position of the robot and when they added blood -like dyes and changed the appearance of the gallbladder and the surrounding tissue.
Last year, Krieger used the system to train a robot to perform three important surgical tasks: manipulation of a needle, lifting body tissue and sewing. But these tasks only lasted a few seconds.
Next, the research team would like to train and test the system in more types of operations and expand its skills in order to carry out complete autonomous operation.
“This work is a big leap from previous efforts, since it is concerned with some of the fundamental obstacles to the provision of autonomous surgical robots in the real world,” said the leading author Ji Woong “Brian” Kim, former post -doctoral research researcher to John’s Hopkins, who is now at Stanford University. “Our work shows that AI models can be made reliably enough for surgical autonomy-etwas that once felt far away, but is now proven to be viable.”
The results were published in the journal Science Robotics.
