Researchers use imitation learning to train surgical robots

Robots are already making some surgical procedures faster and more uniform while improving patient outcomes. However, they haven’t reached their full potential, which is performing fully autonomous surgeries.

Currently, surgical robots must be programmed with each individual move required during a medical procedure. This is a timely process that could still leave the robot unprepared for unplanned circumstances.

Researchers at Johns Hopkins University and Stanford University are hoping to change this. The team used imitation learning to train Intuitive Surgical’s da Vinci robot to perform three fundamental tasks: manipulating a needle, lifting body tissue, and suturing. In each case, the robot trained on the team’s model performed the same surgical procedures as skillfully as human doctors.

The team’s model combines imitation learning with the same machine learning architecture that underpins ChatGPT. However, instead of working with words and text like ChatGPT does, this model works with kinematics.

The researchers fed their model hundreds of videos recorded from wrist cameras placed on the arms of da Vinci robots during surgical procedures. These videos, recorded by surgeons all over the world, are used for post-operative analysis and then archived.

Nearly 7,000 da Vinci robots are used worldwide, and more than 50,000 surgeons are trained on the system, creating a large archive of data for robots to “imitate.”

The researchers said the key was training the model to perform relative movements rather than absolute actions, which are inaccurate.

The model could be used to quickly train surgical robots to perform any type of surgical procedure, the researchers said. The team is now using imitation learning to train a robot to perform not just small surgical tasks but a full operation.

Explore the RBR50 Robotics Innovation Awards 2025.