Professor Zhang Li’s team develops AI-driven magnetic tracking technology to advance clinical use of miniature medical robots

A research team led by Professor Zhang Li from the Department of Mechanical and Automation Engineering has developed an artificial intelligence (AI)-driven magnetic tracking technology. This innovation overcomes a long-standing global bottleneck in real-time micro-robot tracking, enabling the real-time localisation and closed-loop navigation of millimetre-scale wireless magnetic robots during magnetic actuation, paving the way for ultra-precise, minimally invasive therapies.

The research findings have been published in Science Robotics, the world’s leading and most prestigious journal in the field of robotics. This major breakthrough enables tiny wireless robots to be guided safely and precisely through confined and hard-to-access anatomical regions, including blood vessels, bile ducts and the spinal subarachnoid space, and could help advance their clinical translation for minimally invasive diagnosis and therapy. Notably, this study marks the sixth research paper by Professor Zhang’s team published to be published in the journal, underscoring CUHK’s sustained leadership in miniature and medical robotics.

Overcoming the “blind navigation” challenge in micro-robotics

Minimally invasive procedures play a critical role in modern healthcare, enabling diagnosis and treatment in hard-to-reach regions of the body while reducing patient trauma. Miniature wireless magnetic robots offer immense potential for next-generation minimally invasive medicine because they can be remotely guided by magnetic fields into confined anatomical environments that are difficult to reach using conventional medical devices or without open surgery.

Despite their potential, however, the clinical translation of miniature magnetic robots has been hindered by the difficulty of determining their precise location inside the body while they are being magnetically driven. Reliable real-time tracking is essential for guiding them safely to target sites. However, existing tracking technologies are difficult to apply because of the robots’ miniature size, wireless nature and the strong magnetic fields required for actuation.

To overcome this “blind navigation” challenge, the CUHK research team engineered an AI-driven magnetic tracking platform that enables real-time localisation of millimetre-scale wireless magnetic robots during magnetic actuation. The system achieved millimetre-level localisation accuracy at penetration depths of up to 10 centimetres inside biological tissue. The team further integrated the real-time tracking information into the actuation system to enable closed-loop navigation.

The technology was further validated in a large animal model, where miniature magnetic robots were navigated through the spinal subarachnoid space and bile duct of a goat. The experiments were conducted in an unshielded clinical environment with various sources of magnetic interference, and the estimated trajectories showed close agreement with X-ray observations. These results demonstrate robust in vivo closed-loop navigation of miniature magnetic robots in clinically relevant environments and represent an important step towards their clinical translation.

Professor Zhang Li said: “One of the key barriers to the clinical translation of miniature medical robots is the inability to track them accurately in real time while they are being actuated inside the body. Our platform introduces an AI-based approach to suppress actuation-induced interference, enabling reliable closed-loop navigation of millimetre-scale robots in clinically relevant environments. This brings miniature medical robots a major step closer to practical clinical applications in minimally invasive diagnosis and treatment.”

This milestone reflects CUHK’s robust innovation ecosystem and multi-institutional support. This study was supported by the Hong Kong Jockey Club Charities Trust through the JC STEM Early Career Research Fellowship, the Croucher Foundation, the Research Grants Council of Hong Kong (RGC), as well as the SIAT–CUHK Joint Laboratory of Robotics and Intelligent Systems and the Multi-scale Medical Robotics Center (MRC) under the InnoHK research platform.