Revolutionizing Surgical Robotics with Autonomous Intelligence

A pioneering milestone in the realm of Robot-Assisted Surgery (RAS) has been reached in Hong Kong. On August 5th, The Chinese University of Hong Kong revealed a collaborative breakthrough with Cornerstone Robotics, a global leader and unicorn in surgical robotics innovation. Together, they unveiled the world’s first clinical validation of autonomous surgery performed by an intelligent robotic system. This was accomplished using Cornerstone Robotics’ Sentire® Endoscopic Surgical System, an advanced platform designed for multitasking surgical automation powered by embodied artificial intelligence.

Published Research Validates Autonomous Surgical Intelligence

The study, titled “Surgical Embodied Intelligence for Generalized Task Autonomy in Laparoscopic Robot-assisted Surgery,” was featured in Science Robotics, a leading journal in robotics research boasting a five-year impact factor of 32.9. This publication marks the first successful in-vivo animal trial of a Chinese-developed surgical robot equipped with AI-driven embodied intelligence. The results demonstrate the system’s ability to autonomously perform multiple surgical tasks while collaborating seamlessly with human operators, highlighting the clinical viability and adaptability of the underlying algorithms. This breakthrough paves the way for smarter, more accessible robotic surgery solutions worldwide.

Introducing the VPPV Framework: A New Era in Surgical Automation

Central to this innovation is the novel VPPV framework, which integrates Visual Parsing, Perceptual Regression, Policy Learning, and Visual Servoing Control. This hybrid approach synergizes robust foundational AI models with reinforcement learning and classical control theory, enabling unprecedented stability, flexibility, and repeatability in autonomous surgical tasks. Unlike previous systems reliant on rigid pre-programmed rules, this vision-based intelligence framework supports full perception, multitasking, precise closed-loop control, and dynamic human-robot interaction. It significantly lowers barriers to research, clinical adoption, and sets a new benchmark for next-generation intelligent surgical assistants.

Training and Simulation with SurRoL: Bridging Virtual and Real Worlds

The research team utilized SurRoL, an independently developed open-source surgical embodied intelligence simulator, to train multiple AI models capable of generalized task autonomy. SurRoL combines multimodal sensory inputs, a physics-based simulation engine, and reinforcement learning algorithms to enable zero-shot transfer from simulation to real-world surgical environments. This comprehensive platform includes a robot-assisted surgical control interface, simulation engine, and learning modules, accelerating the development and validation of autonomous surgical capabilities.

Robust Ex Vivo and In Vivo Testing on the Sentire® System

To replicate authentic surgical conditions, the team conducted extensive experiments on ex vivo porcine stomach tissue preserving critical arteries, as well as on a live ~30kg pig. These trials employed the Sentire® Endoscopic Surgical System, fully engineered by Cornerstone Robotics. During in vivo testing, the system operated under a novel supervised autonomy safety mode, relying solely on its high-definition stereoscopic endoscope without external sensors. The robot autonomously executed complex tasks such as gauze retrieval, soft tissue manipulation, and vascular clipping, demonstrating remarkable adaptability, control precision, and clinical translation potential. This validates the feasibility of embodied intelligence-driven human-robot collaboration in surgical settings.

Cornerstone Robotics’ Vertically Integrated, AI-Optimized Platform

Dr. Jerry Wang, Co-Founder, CTO, and COO of Cornerstone Robotics, emphasized the comprehensive nature of the Sentire® system: “Our platform is built entirely in-house, encompassing mechanical design, electronics, software architecture, robotic algorithms, imaging, and advanced energy management. This commercial-grade system delivers the precision and robustness essential for deploying AI autonomy in complex surgical environments.”

He further noted, “Our commitment to full-stack independent development and vertical integration ensures exceptional scalability and AI-readiness. This architecture facilitates rapid AI algorithm deployment and flexible adaptation to diverse surgical procedures, laying a strong foundation for the future of intelligent robotic surgery.”

From Research to Clinical Reality: Accelerating AI-Driven Surgical Innovation

Cornerstone Robotics collaborated closely with CUHK researchers to implement and validate the autonomous algorithms on the Sentire® system, marking a significant step toward clinical translation. This achievement not only confirms the robustness and compatibility of Cornerstone’s open-source infrastructure but also accelerates the integration of AI-powered surgical assistance into real-world medical practice.

Professor Samuel Au, Founder and CEO of Cornerstone Robotics, remarked, “The fusion of academic research and industry expertise is crucial for advancing medical technology. This collaboration demonstrates the practical application of embodied AI on a commercial robotic platform and highlights its transformative potential in surgical automation. We are dedicated to partnering with global researchers and clinicians to explore new frontiers in task automation, human-robot synergy, and intelligent system monitoring, driving the clinical adoption of smart surgical robots.”

About Cornerstone Robotics Limited

Established in 2019, Cornerstone Robotics (CSR) is a unicorn enterprise at the forefront of medical innovation, committed to enhancing global healthcare through advanced robotic systems. With a multidisciplinary team spanning Hong Kong, Shenzhen, Beijing, Shanghai, Boston, and Portsmouth, CSR is rapidly advancing robotic solutions for soft tissue and other surgical specialties. Their mission is to make high-quality surgical care more accessible and efficient worldwide.

For more information, visit https://en.csrbtx.com/.