Robot-patient registration for optical tracker-free robotic fracture reduction surgery

Area of Science:

• Medical Robotics
• Surgical Navigation
• Image-Guided Surgery

Background:

• Robotic surgery for fracture reduction requires accurate robot-patient registration for navigation.
• Current methods often rely on optical trackers, which can be cumbersome.
• Estimating the 3D relationship between the robot and patient's bones is crucial for simulating fracture states.

Purpose of the Study:

• To develop an optical-tracker-free robot-patient registration method for robotic fracture reduction.
• To utilize X-ray imaging for precise registration in a robotic surgical system.
• To enable accurate simulation of real-world fracture states in virtual navigation systems.

Main Methods:

• A two-step registration process: initial and refined registration using particle swarm optimization.
• Employing bidirectional X-ray images to minimize cross-reprojection error.
• Developing and integrating attachable robot features for clear extraction from X-ray images, overcoming interference issues.

Main Results:

• Phantom experiments yielded average translational and rotational errors of 1.88 mm and 2.45°.
• Ex vivo experiments with a caprine cadaver showed average errors of 2.64 mm and 3.32°.
• The results confirm the effectiveness of the proposed robot-patient registration method.

Conclusions:

• The method accurately estimates the 3D bone relationship using 2D X-ray images.
• This enables precise simulation in virtual reality for surgical navigation.
• The registration method is expected to aid successful bone fracture treatment in image-guided robotic surgery.