基于卡尔曼滤波的髓内钉锁孔操作标定方法研究

Study on a Calibration Method Based on Kallman Filter for Locking of Intramedullary Nail

在机器人辅助正骨手术中,髓内钉的锁孔操作需要很高的定位精确度,因此需要一种相应的高精确度系统标定方法来满足定位要求.本文利用卡尔曼滤波的间接法标定法,其跳过对定位盒顶点坐标的求取,将标定盒三个顶点的坐标作为状态参量,将机器人由拍照位姿到钻孔位姿之间的位姿变换参数当作观测参量.并在此基础上推导出状态参量与观测参量之间的映射关系,对一系列观测参量的测量,利用卡尔曼滤波算法对状态参量进行滤波估计,最后间接求取转换矩阵.实验证明,这种方法的精确度不会因为标定试验次数增多而导致标定盒磨损,重复精确度较高.

In robot-assisted orthopedic surgery, the locking of intramedullary nail needs to be operated with high position accuracy. So a calibration method with high accuracy is also needed. In this paper a method based on Kallman filter is presented. It avoideds calculating the coordinate values of vertex on the locating box but treats them as the state variables. The parameters of transformation from the pose of taking photos to the pose of drilling are treated as observation variables. The mapping relationship between the state variables and the observation variables are deduced. With the measuring of the observation variables and using Kalman filter to estimate the real states, the transformation matrix can be obtained indirectly. The experiment results show that the calibration accuracy of this method will not be affected by the abrasion of locating box due to the multiple calibrations, and with a high repeatability.