Kyung Hee University Begins Development of Next-Generation Autonomous Intervention Robot

Pioneering Automation of Neurovascular Procedures
Through AI and Robotics Integration

Kyung Hee University announced on September 20 that the research team led by Professor Hyo-Suk Hwang of the Department of Software Convergence has been ultimately selected for the advanced stage of the 2026 Basic Research Laboratory Support Project (BRL), a newly funded project supported by the Ministry of Science and ICT and the National Research Foundation of Korea.

Hwang Hyoseok (left), Professor of Software Convergence at Kyung Hee University; Kim Jongwoo, Professor of Mechanical Engineering at Kyung Hee University; Cho Myeonga, Professor of Software Convergence at Kyung Hee University; and Yoon Wonki, Professor at Korea University College of Medicine. Kyung Hee University

The research team will operate a "Research Laboratory for AI Autonomous Neurovascular Intervention Robots" for three years until June 2029, with a total research grant of 1.5 billion won.

This research project will be led by Professor Hyo-Suk Hwang as the principal investigator, with Professor Jongwoo Kim from the Department of Mechanical Engineering, Professor Myung-A Cho from the Department of Software Convergence, both at Kyung Hee University, and Professor Won-Gi Yoon from Korea University College of Medicine joining as co-researchers.

In addition, the University of Toronto in Canada and SickKids (PCIGITI Center), a world-renowned children's hospital, will participate in the project, making it a globally competitive international joint research endeavor.

The main objective of this research is to automate catheter procedures for stroke and cerebral aneurysms—traditionally performed manually by physicians—through autonomous decision-making powered by AI and precise robotic control.

Conventional endovascular interventions are highly challenging because physicians must navigate complex three-dimensional vessels using only cross-sectional imaging. To overcome these limitations, the research team plans to establish an integrated system that includes: ▲ active variable stiffness "Hard-Soft Hybrid" continuum robots ▲ medical twins for patient-specific vascular reconstruction ▲ reinforcement learning-based AI autonomous navigation ▲ high-fidelity phantom verification systems.

Notably, the system applies a hydraulic multi-chamber technology to dynamically adjust the robot's localized stiffness in real time. The team will also introduce a three-dimensional vascular map generated from a single MRA image and utilize generative AI-driven data augmentation technology to immediately address the paucity of training data for AI learning.

Through these innovations, the team aims to develop a fully integrated system in which AI and robotics not only provide assistance but autonomously execute the entire procedure—from pathfinding to stent and coil placement.

The research team has set specific benchmarks: a 20% reduction in procedure time, a target success rate of 95% or higher for reaching the intended point, and a 30% improvement in procedural safety.

If this technology is commercialized, it is expected to significantly enhance the neurovascular emergency response capabilities of regional hospitals experiencing shortages of specialized medical professionals, thereby contributing to closing the gap in healthcare access between regions.

Professor Hyo-Suk Hwang of the Department of Software Convergence at Kyung Hee University stated, "By converging AI and robotic technologies, our core aim is to minimize variations in practitioners’ skill levels so that patients can receive the best treatment possible, anytime and anywhere." He added, "The variable stiffness and AI autonomous control technologies developed through this research are key foundational innovations that can be expanded to many fields, including brain surgery and cardiovascular interventions. As such, we are committed to advancing the localization of high-value-added medical robot components, which have so far been highly dependent on imports."

This content was produced with the assistance of AI translation services.