A Registration Method for Total Knee Arthroplasty Surgical Robot

Yuankun Wei,Ziyu Fu,Debing Zhang,Guangqi Wang,Chun Zhang,Xiang Xie,Xiangyu Li
DOI: https://doi.org/10.1109/itoec53115.2022.9734563
2022-01-01
Abstract:Robot used in TKA (Total Knee Arthroplasty) can significantly improve accuracy. The surgical navigation system is an important part of the robot, which needs to complete the registration. The registration process needs to calculate the transformation matrix of the patient's bone and the virtual model of the patient's bone. This paper proposes a registration method based on Point-to-Plane ICP (Iterative Closest Point). Before the operation, navigation system will extract target point cloud from the virtual model and generate a guiding point set. During the operation, the doctor uses the probe to touch the patient's bone following the guiding point set to generate a spare point cloud. After the registration of spare point cloud and target point cloud, the transformation matrix is given, and then the navigation system would trace patient's bone by reference frame. It requires only 40 points on each bone to produce an accurate result. Experiments show that average time for registration is 0.36s, and the average RMS error is 0.3mm.
What problem does this paper attempt to address?