Edge computing for robotic telesurgery using artificial intelligence

A surgical robotic system using edge computing, comprising:

• a control system configured to:
• cause a robotic arm of a surgical robot to robotically position one or more sensors to capture images of a surgical site;
• generate, by an artificial intelligence (AI) accelerator, a prediction of an adverse surgical event, using a machine learning (ML) model, based on the images of the surgical site received from the one or more sensors, the prediction indicating that insertion of a first implant would touch non-targeted tissue at the surgical site;
• in response to the prediction of the adverse surgical event, transitioning control of the surgical robot to an edge computing system, wherein the edge computing system is configured to:
• generate instructions for controlling the robotic arm to avoid the adverse surgical event during robotic telesurgery, the instructions directing insertion of a second implant different from the first implant to avoid touching the non-targeted tissue by the second implant,
• wherein the AI accelerator of the edge computing system generates the prediction and the instructions on-the-edge while avoiding concurrent data transfer to a computer system over a network;
• move, by robotic joints of the robotic arm, the robotic arm to perform robotic telesurgery, wherein the robotic arm is communicably coupled to the edge computing system;
• position the robotic arm in a configuration based on the instructions, the configuration defined by a spatial position and an orientation;
• and
• perform at least one action to avoid the adverse surgical event when the robotic arm is positioned in the configuration;
• determine the adverse surgical event was avoided by the surgical robot;
• and
• in response to determining the adverse surgical event was avoided, transition control of the surgical robot from the edge computing system to the control system.

A system comprising:

• one or more computer processors;
• and
• a non-transitory computer-readable storage medium storing computer instructions, which when executed by the one or more computer processors, cause the system to:
• cause, by a control system, a surgical robot to robotically position one or more sensors to capture images of a surgical site, wherein the surgical robot comprises a robotic arm;
• generate, by an artificial intelligence (AI) accelerator of the control system, a prediction of an adverse surgical event, using a machine learning (ML) model, based on images of a surgical site, the prediction indicating that insertion of a first implant would touch non-targeted tissue at the surgical site;
• in response to the prediction of the adverse surgical event, transitioning control of the surgical robot to an edge computing system, wherein the edge computing system performs steps comprising:
• generate, by the AI accelerator, instructions for controlling the surgical robot via the edge computing system to avoid the adverse surgical event during robotic telesurgery, the instructions directing insertion of a second implant different from the first implant to avoid touching the non-targeted tissue by the second implant,
• wherein the AI accelerator of the edge computing system generates the prediction and the instructions on-the-edge while avoiding concurrent data transfer to a computer system over a network;
• move, by robotic joints of the robotic arm, the robotic arm to perform the robotic telesurgery, wherein the robotic arm is communicably coupled to the edge computing system;
• position, by the surgical robot, the robotic arm in a configuration based on the instructions, the configuration defined by a spatial position and an orientation;
• and
• perform, by the surgical robot, at least one action to avoid the adverse surgical event when the robotic arm is positioned in the configuration;
• determine the adverse surgical event was avoided by the surgical robot;
• and
• in response to determining the adverse surgical event was avoided, transition control of the surgical robot from the edge computing system to the control system.