The innovative framework, VPPV Embodied Intelligence, integrates visual parsing, perceptual regression, policy learning, and visual servoing, allowing the robot to automate surgical tasks without the need for additional sensors.

Through training multiple AI frameworks on an open-source simulator, the system has demonstrated advanced capabilities, including full perception, multitask applicability, precise closed-loop control, and effective human-robot cooperation.

The technology aims to assist surgeons by automating routine tasks, which could significantly reduce their workload and enhance surgical efficiency.

Professor Samuel Au, CEO of Cornerstone Robotics, emphasized the importance of collaboration between research and industry in advancing medical technologies, particularly in the realm of surgical automation.

Dr. Jerry Wang, CTO of Cornerstone Robotics, highlighted the fully integrated infrastructure of the Sentire® system, which is crucial for stable AI deployment in dynamic surgical environments.

A groundbreaking study published in Science Robotics has successfully demonstrated the first in vivo animal testing of a Chinese-developed laparoscopic surgical robot, showcasing its AI-powered embodied intelligence and capabilities for multitask autonomy and human-robot collaboration.

This achievement was made possible by Cornerstone Robotics in collaboration with the Chinese University of Hong Kong (CUHK), marking the world's first clinical validation of autonomous surgery with their Sentire® Endoscopic Surgical System.

The Sentire® system is designed to analyze endoscopic images in real-time and employs advanced techniques such as visual foundation models and reinforcement learning to perform various surgical tasks.

The InnoHK Multi-Scale Medical Robotics Center facilitated this research through international collaboration, notably with Johns Hopkins University, supporting high-impact advancements in surgical robotics.

Rigorous testing, including ex vivo and in vivo trials, demonstrated the robot's ability to autonomously perform surgical tasks such as gauze picking and blood vessel clipping while maintaining adaptability and control stability.

Professor Dou Qi led the study, pointing out that traditional surgical automation methods often limited clinical applicability, a challenge that the new vision-based solution successfully addresses.

The open-sourcing of the AI software infrastructure has enabled numerous research institutions worldwide to adopt and further develop this innovative technology.