Robots Trained on Videos Using da Vinci System Achieve Human Surgeon Skill Levels, Self-Correcting Errors

Dec 30, 2024, 06:10 PM

Researchers from Johns Hopkins University and Stanford University have achieved a breakthrough in robotic surgery by training robots to perform surgical tasks autonomously using video observation. The robots were trained through imitation learning, a method where they analyzed videos of surgical procedures performed by human doctors. This approach enabled the robots to perform tasks such as suturing, knot-tying, and manipulating needles with the skill level of human surgeons. Notably, the robots demonstrated the ability to correct their own mistakes, such as picking up dropped needles, without human intervention. The research, which utilized the da Vinci Surgical System, marks a step forward in medical robotics, potentially addressing the shortage of surgeons and reducing medical errors. The findings were presented at the Conference on Robot Learning in Munich. The project, led by Axel Krieger and Ji Woong 'Brian' Kim, also highlighted the use of 7,000 da Vinci robots worldwide. The next phase of the research involves training robots to perform complete surgeries on animal cadavers. While the technology shows promise, concerns remain about the reliability of autonomous robots in handling unpredictable surgical scenarios and their regulatory approval.