脑外科机器人视觉伺服研究

Visual servoing control in robot-assisted stereotactic neurosurgery

定位精度是脑外科机器人系统一个最重要的指标.提出了利用视觉伺服提高脑外科机器人系统定位精度的方法.根据机器人辅助脑外科手术的特点,选择基于位置的末端点闭环的视觉伺服方式,提出了一种利用机器人运动信息对机器人的特征点进行视觉跟踪定位的稳定可靠的方法.针对病人体内病灶点在摄像机图像中不可见的问题,推导了脑外科机器人系统的视觉伺服控制律.采用视觉伺服的方法后,系统定位精度有了很大提高,系统定位误差主要由靶点映射误差引起,受机器人绝对定位误差的影响不大.手术过程模拟和定位精度测试验证了该方法的有效性.

Positioning accuracy is one of the most important specifications in robot-assisted stereotactic neurosurgery. A visual servoing control method in robot-assisted stereotactic neurosurgery was presented to improve the overall positioning accuracy of the robot system. The position-based endpoint closed-loop visual servoing architecture was selected according to the characteristic of robot-assisted stereotactic neurosurgery. A visual tracking method utilizing the robot motion information was developed to locate the robot end-effector in the cameras images in real time. The visual servoing control law for the robot system was derived aiming at the problem that the tumor inside the patient＇s head was unseen in the cameras images. With the visual servoing control method, the positioning error of the robot system was mainly caused by the target registration process. The effect of the robot absolute positioning error was eliminated and the overall positioning accuracy of the robot system was improved. Simulation of the clinical process and positioning accuracy test proved the validity of the method.